utterance
stringlengths
1
128
goal_semantics
stringclasses
127 values
base_utterance
stringclasses
139 values
json_semantics
stringclasses
127 values
intent
stringclasses
3 values
central_proposition
stringclasses
38 values
supplemental_semantics
stringclasses
125 values
stylizer
stringclasses
8 values
augmenter
float64
actions
stringclasses
1 value
properties
stringclasses
1 value
prompt_template
stringclasses
1 value
text
stringlengths
596
770
input
stringlengths
1
128
output
stringclasses
127 values
instruction
stringclasses
1 value
evan owns that hair drier, y'all.
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance evan owns that hair drier, y'all. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
evan owns that hair drier, y'all.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
uh, the hair, um, drier is, uh, evan's
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance uh, the hair, um, drier is, uh, evan's ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
uh, the hair, um, drier is, uh, evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
so, the hair drier, i mean, it's evan's
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance so, the hair drier, i mean, it's evan's ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
so, the hair drier, i mean, it's evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the blow dryer belongs to evan
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the blow dryer belongs to evan ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the blow dryer belongs to evan
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
evan owns the hairdryer
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance evan owns the hairdryer ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
evan owns the hairdryer
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the fair trial is evan's
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the fair trial is evan's ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the fair trial is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the hair dryer is evans
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the hair dryer is evans ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the hair dryer is evans
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the hair drier is even vans
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the hair drier is even vans ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the hair drier is even vans
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the hair drier is belonging to john, no wait, i mean it's evan's
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the hair drier is belonging to john, no wait, i mean it's evan's ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the hair drier is belonging to john, no wait, i mean it's evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the hair drier is kate's, sorry my mistake, it's actually evan's
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the hair drier is kate's, sorry my mistake, it's actually evan's ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the hair drier is kate's, sorry my mistake, it's actually evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the hair drier is evan's
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{hairdrier(VAR0),DEFINITE(VAR0)})
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['hairdrier(VAR0)']
style:NoStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the hair drier is evan's ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
the hair drier is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['hairdrier(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
it appears that the bicycle is of a yellow hue.
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance it appears that the bicycle is of a yellow hue. ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
it appears that the bicycle is of a yellow hue.
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
i think you may find that the bicycle has a yellow color.
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance i think you may find that the bicycle has a yellow color. ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
i think you may find that the bicycle has a yellow color.
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
bike, it is of yellow color
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance bike, it is of yellow color ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
bike, it is of yellow color
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the color of bicycle, it is yellow yes?
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the color of bicycle, it is yellow yes? ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the color of bicycle, it is yellow yes?
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bike's yellow, dude
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bike's yellow, dude ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bike's yellow, dude
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
got a yellow bike here
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance got a yellow bike here ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
got a yellow bike here
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
um, the bicycle, it's, uh, yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance um, the bicycle, it's, uh, yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
um, the bicycle, it's, uh, yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
so, the bicycle, you know, it's like, yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance so, the bicycle, you know, it's like, yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
so, the bicycle, you know, it's like, yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bike is colored in gold
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bike is colored in gold ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bike is colored in gold
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the cycle has a hue of canary
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the cycle has a hue of canary ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the cycle has a hue of canary
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bicycle is mellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bicycle is mellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bicycle is mellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bicycle yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bicycle yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bicycle yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bicycle is is yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bicycle is is yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bicycle is is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bicycle is blue, oh wait, no, it's actually yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bicycle is blue, oh wait, no, it's actually yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bicycle is blue, oh wait, no, it's actually yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bike, i mean, the bicycle is....um...it's yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bike, i mean, the bicycle is....um...it's yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bike, i mean, the bicycle is....um...it's yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the bicycle is yellow
STATEMENT(tyler,self:agent,propertyOf(VAR0,yellow),{bicycle(VAR0),DEFINITE(VAR0)})
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
STATEMENT
propertyOf(VAR0,yellow)
['bicycle(VAR0)']
style:NoStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the bicycle is yellow ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
the bicycle is yellow
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,yellow)', 'supplemental_semantics': ['bicycle(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
i believe the skis belong to chris.
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance i believe the skis belong to chris. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
i believe the skis belong to chris.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
it seems that the skis are in chris' possession.
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance it seems that the skis are in chris' possession. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
it seems that the skis are in chris' possession.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis, they are belonging to chris
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis, they are belonging to chris ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis, they are belonging to chris
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
chris, he is the owner of the skis
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance chris, he is the owner of the skis ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
chris, he is the owner of the skis
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
w
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance w ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
w
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
?
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance ? ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
?
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
uh, the skis, um, they are, um, chris'
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance uh, the skis, um, they are, um, chris' ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
uh, the skis, um, they are, um, chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
so, um, the skis are, er, chris'
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance so, um, the skis are, er, chris' ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
so, um, the skis are, er, chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis belong to chris
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis belong to chris ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis belong to chris
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
chris is the owner of the skis
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance chris is the owner of the skis ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
chris is the owner of the skis
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the keys are chris
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the keys are chris ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the keys are chris
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis are chairs
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis are chairs ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis are chairs
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis are
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis are ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis are
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis are, uh i mean, the skis are chris'
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis are, uh i mean, the skis are chris' ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis are, uh i mean, the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis are mine, no, i mean, the skis are chris'
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis are mine, no, i mean, the skis are chris' ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis are mine, no, i mean, the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the skis are chris'
STATEMENT(tyler,self:agent,propertyOf(VAR0,chris'),{skis(VAR0),DEFINITE(VAR0)})
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
STATEMENT
propertyOf(VAR0,chris')
['skis(VAR0)']
style:NoStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the skis are chris' ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
the skis are chris'
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,chris')", 'supplemental_semantics': ['skis(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
i find the tennis racket to be a bit on the heavier side.
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance i find the tennis racket to be a bit on the heavier side. ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
i find the tennis racket to be a bit on the heavier side.
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
you might find the tennis racket somewhat heavier than usual.
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance you might find the tennis racket somewhat heavier than usual. ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
you might find the tennis racket somewhat heavier than usual.
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
tennis bat is much weight
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance tennis bat is much weight ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
tennis bat is much weight
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis hit thing, it have much heaviness.
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis hit thing, it have much heaviness. ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis hit thing, it have much heaviness.
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
this tennis racket's got some weight to it!
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance this tennis racket's got some weight to it! ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
this tennis racket's got some weight to it!
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
dude, this tennis racket is so heavy!
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance dude, this tennis racket is so heavy! ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
dude, this tennis racket is so heavy!
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
uh, the tennis, um, racket is, you know, heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance uh, the tennis, um, racket is, you know, heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
uh, the tennis, um, racket is, you know, heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis, um, racket, it's like, heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis, um, racket, it's like, heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis, um, racket, it's like, heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis bat is weighty
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis bat is weighty ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis bat is weighty
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the racquet for tennis feels hefty
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the racquet for tennis feels hefty ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the racquet for tennis feels hefty
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis rocket is heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis rocket is heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis rocket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis is heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis is heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis racket is heavy and expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis racket is heavy and expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis racket is heavy and expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis ball is, no wait, the tennis racket is heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis ball is, no wait, the tennis racket is heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis ball is, no wait, the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis racket is light, actually, i got that wrong, the tennis racket is heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis racket is light, actually, i got that wrong, the tennis racket is heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis racket is light, actually, i got that wrong, the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the tennis racket is heavy
STATEMENT(tyler,self:agent,propertyOf(VAR0,heavy),{tennisracket(VAR0),DEFINITE(VAR0)})
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
STATEMENT
propertyOf(VAR0,heavy)
['tennisracket(VAR0)']
style:NoStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the tennis racket is heavy ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
the tennis racket is heavy
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,heavy)', 'supplemental_semantics': ['tennisracket(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
it appears to me that the train is of a blue color
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance it appears to me that the train is of a blue color ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
it appears to me that the train is of a blue color
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
i believe one might observe that the train is indeed, blue in color.
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance i believe one might observe that the train is indeed, blue in color. ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
i believe one might observe that the train is indeed, blue in color.
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
blue is the train
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance blue is the train ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
blue is the train
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
train, it is of blue color
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance train, it is of blue color ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
train, it is of blue color
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
!
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance ! ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
!
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
uh, the train, um, it's blue
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance uh, the train, um, it's blue ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
uh, the train, um, it's blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the, um, train is, uh, blue
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the, um, train is, uh, blue ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the, um, train is, uh, blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the locomotive is azure
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the locomotive is azure ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the locomotive is azure
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the railway vehicle is cobalt
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the railway vehicle is cobalt ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the railway vehicle is cobalt
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the drain is blue
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the drain is blue ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the drain is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the train is
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the train is ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the train is
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the train is blue and
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the train is blue and ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the train is blue and
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the car is blue, wait, no, i meant the train is blue
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the car is blue, wait, no, i meant the train is blue ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the car is blue, wait, no, i meant the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the plane is blue, my apologies, i meant the train is blue
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the plane is blue, my apologies, i meant the train is blue ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the plane is blue, my apologies, i meant the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the train is blue
STATEMENT(tyler,self:agent,propertyOf(VAR0,blue),{train(VAR0),DEFINITE(VAR0)})
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
STATEMENT
propertyOf(VAR0,blue)
['train(VAR0)']
style:NoStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the train is blue ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
the train is blue
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,blue)', 'supplemental_semantics': ['train(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy bear carries a rather high price tag, don't you think?
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy bear carries a rather high price tag, don't you think? ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy bear carries a rather high price tag, don't you think?
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
would you agree that the cost of the teddy bear is on the higher side?
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance would you agree that the cost of the teddy bear is on the higher side? ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
would you agree that the cost of the teddy bear is on the higher side?
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
bear of teddy, it is costing much
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance bear of teddy, it is costing much ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
bear of teddy, it is costing much
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy, it is having high price
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy, it is having high price ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy, it is having high price
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
that teddy bear costs an arm and a leg
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance that teddy bear costs an arm and a leg ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
that teddy bear costs an arm and a leg
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
damn, this teddy bear ain't cheap!
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance damn, this teddy bear ain't cheap! ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
damn, this teddy bear ain't cheap!
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
uh, the teddy, um, bear is, like, expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance uh, the teddy, um, bear is, like, expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
uh, the teddy, um, bear is, like, expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
so, the teddy bear is, uh, expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance so, the teddy bear is, uh, expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
so, the teddy bear is, uh, expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the plush toy is costly
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the plush toy is costly ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the plush toy is costly
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the stuffed animal is high-priced
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the stuffed animal is high-priced ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the stuffed animal is high-priced
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy beer is expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy beer is expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy beer is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy bear expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy bear expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy bear expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy bear is really expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:ASRStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy bear is really expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy bear is really expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the toy car is, wait no, the teddy bear is expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the toy car is, wait no, the teddy bear is expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the toy car is, wait no, the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy bear is cheap, no that's not right, the teddy bear is expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:CorrectionStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy bear is cheap, no that's not right, the teddy bear is expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy bear is cheap, no that's not right, the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the teddy bear is expensive
STATEMENT(tyler,self:agent,propertyOf(VAR0,expensive),{teddybear(VAR0),DEFINITE(VAR0)})
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
STATEMENT
propertyOf(VAR0,expensive)
['teddybear(VAR0)']
style:NoStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the teddy bear is expensive ### JSON: {'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
the teddy bear is expensive
{'intent': 'STATEMENT', 'central_proposition': 'propertyOf(VAR0,expensive)', 'supplemental_semantics': ['teddybear(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
i believe the handbag belongs to evan.
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance i believe the handbag belongs to evan. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
i believe the handbag belongs to evan.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
it appears that the handbag is actually evan's.
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:DirectnessStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance it appears that the handbag is actually evan's. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
it appears that the handbag is actually evan's.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
evan, he own the handbag
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance evan, he own the handbag ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
evan, he own the handbag
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
handbag, it belong evan.
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:FamiliarityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance handbag, it belong evan. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
handbag, it belong evan.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
that handbag is evan's yo
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance that handbag is evan's yo ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
that handbag is evan's yo
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
evan's got dibs on the handbag
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:FormalityStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance evan's got dibs on the handbag ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
evan's got dibs on the handbag
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
uh, the handbag, it's um, evan's.
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance uh, the handbag, it's um, evan's. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
uh, the handbag, it's um, evan's.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
so, the hand...handbag, it's, erm, evan's.
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:DisfluencyStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance so, the hand...handbag, it's, erm, evan's. ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
so, the hand...handbag, it's, erm, evan's.
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
the purse belongs to evan
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance the purse belongs to evan ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
the purse belongs to evan
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}
evan owns the purse
STATEMENT(tyler,self:agent,propertyOf(VAR0,evan's),{handbag(VAR0),DEFINITE(VAR0)})
the handbag is evan's
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
STATEMENT
propertyOf(VAR0,evan's)
['handbag(VAR0)']
style:WordChoiceStyleAugmenter
null
['putabove', 'handleAck', 'take', 'stopVisualSearch', 'putoutside', 'putagainst', 'putrightof', 'putamong', 'getTime', 'putbehind', 'putbelow', 'lookForObject', 'initSearchesDemo', 'puton', 'putinside', 'putallover', 'startVisualSearch', 'putalong', 'findObject', 'translateLastGoal', 'initSearches', 'putleftof', 'getCurrGoals', 'clearrelations', 'findGraspableObject', 'handleGreeting', 'putbeside', 'putinfrontof']
['doit', 'dothis', 'dothat', 'that', 'this', 'it', 'that', 'thing', 'those', 'they', 'these', 'this', 'physobj', 'at', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'trafficlight', 'firehydrant', 'stopsign', 'parkingmeter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball', 'kite', 'baseballbat', 'baseballglove', 'skateboard', 'surfboard', 'tennisracket', 'bottle', 'wineglass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hotdog', 'pizza', 'donut', 'cake', 'chair', 'couch', 'pottedplant', 'bed', 'diningtable', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cellphone', 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddybear', 'hairdrier', 'toothbrush', 'medicine', 'bandagebox', 'painkillers', 'mug', 'bag', 'glass', 'camera', 'box', 'blue', 'red', 'yellow', 'heavy', "evan's", "vasanth's", "chris'", 'pretty', 'scary', 'expensive', 'orange', 'purple', 'green', 'cheap', 'light', "marlow's", "ravenna's", 'it', 'this', 'that', 'thing', 'those', 'they', 'these']
### instruction {instruction} ### example {example} ### utterance {input} ### JSON: {output}
### instruction Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. ### example Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }} ### utterance evan owns the purse ### JSON: {'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
evan owns the purse
{'intent': 'STATEMENT', 'central_proposition': "propertyOf(VAR0,evan's)", 'supplemental_semantics': ['handbag(VAR0)']}
Given an utterance, extract a semantic parse of the utterance and respond with the parse in a perfect JSON format. Here is an example of a parse for an utterance. utterance: put the potted plant outside of the skis JSON: {{ "intent": "INSTRUCT", "central_proposition": "putoutside(self:agent,VAR0,VAR1)", "supplemental_semantics": [[ "pottedplant(VAR0)", "skis(VAR1)", "DEFINITE(VAR0)", "DEFINITE(VAR1)" ]] }}