Davide1999 commited on
Commit
231944f
1 Parent(s): c8f5383

Add SetFit ABSA model

Browse files
1_Pooling/config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 768,
3
+ "pooling_mode_cls_token": false,
4
+ "pooling_mode_mean_tokens": true,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false,
9
+ "include_prompt": true
10
+ }
README.md ADDED
@@ -0,0 +1,219 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: sentence-transformers/paraphrase-mpnet-base-v2
3
+ library_name: setfit
4
+ metrics:
5
+ - accuracy
6
+ pipeline_tag: text-classification
7
+ tags:
8
+ - setfit
9
+ - absa
10
+ - sentence-transformers
11
+ - text-classification
12
+ - generated_from_setfit_trainer
13
+ widget:
14
+ - text: bar:After really enjoying ourselves at the bar we sat down at a table and
15
+ had dinner.
16
+ - text: interior decor:this little place has a cute interior decor and affordable
17
+ city prices.
18
+ - text: cuisine:The cuisine from what I've gathered is authentic Taiwanese, though
19
+ its very different from what I've been accustomed to in Taipei.
20
+ - text: dining:Go here for a romantic dinner but not for an all out wow dining experience.
21
+ - text: Taipei:The cuisine from what I've gathered is authentic Taiwanese, though
22
+ its very different from what I've been accustomed to in Taipei.
23
+ inference: false
24
+ ---
25
+
26
+ # SetFit Aspect Model with sentence-transformers/paraphrase-mpnet-base-v2
27
+
28
+ This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Aspect Based Sentiment Analysis (ABSA). This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification. In particular, this model is in charge of filtering aspect span candidates.
29
+
30
+ The model has been trained using an efficient few-shot learning technique that involves:
31
+
32
+ 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
33
+ 2. Training a classification head with features from the fine-tuned Sentence Transformer.
34
+
35
+ This model was trained within the context of a larger system for ABSA, which looks like so:
36
+
37
+ 1. Use a spaCy model to select possible aspect span candidates.
38
+ 2. **Use this SetFit model to filter these possible aspect span candidates.**
39
+ 3. Use a SetFit model to classify the filtered aspect span candidates.
40
+
41
+ ## Model Details
42
+
43
+ ### Model Description
44
+ - **Model Type:** SetFit
45
+ - **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2)
46
+ - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
47
+ - **spaCy Model:** en_core_web_lg
48
+ - **SetFitABSA Aspect Model:** [Davide1999/setfit-absa-paraphrase-mpnet-base-v2-restaurants-aspect](https://huggingface.co/Davide1999/setfit-absa-paraphrase-mpnet-base-v2-restaurants-aspect)
49
+ - **SetFitABSA Polarity Model:** [Davide1999/setfit-absa-paraphrase-mpnet-base-v2-restaurants-polarity](https://huggingface.co/Davide1999/setfit-absa-paraphrase-mpnet-base-v2-restaurants-polarity)
50
+ - **Maximum Sequence Length:** 512 tokens
51
+ - **Number of Classes:** 2 classes
52
+ <!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
53
+ <!-- - **Language:** Unknown -->
54
+ <!-- - **License:** Unknown -->
55
+
56
+ ### Model Sources
57
+
58
+ - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
59
+ - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
60
+ - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
61
+
62
+ ### Model Labels
63
+ | Label | Examples |
64
+ |:----------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
65
+ | aspect | <ul><li>'staff:But the staff was so horrible to us.'</li><li>"food:To be completely fair, the only redeeming factor was the food, which was above average, but couldn't make up for all the other deficiencies of Teodora."</li><li>"food:The food is uniformly exceptional, with a very capable kitchen which will proudly whip up whatever you feel like eating, whether it's on the menu or not."</li></ul> |
66
+ | no aspect | <ul><li>"factor:To be completely fair, the only redeeming factor was the food, which was above average, but couldn't make up for all the other deficiencies of Teodora."</li><li>"deficiencies:To be completely fair, the only redeeming factor was the food, which was above average, but couldn't make up for all the other deficiencies of Teodora."</li><li>"Teodora:To be completely fair, the only redeeming factor was the food, which was above average, but couldn't make up for all the other deficiencies of Teodora."</li></ul> |
67
+
68
+ ## Uses
69
+
70
+ ### Direct Use for Inference
71
+
72
+ First install the SetFit library:
73
+
74
+ ```bash
75
+ pip install setfit
76
+ ```
77
+
78
+ Then you can load this model and run inference.
79
+
80
+ ```python
81
+ from setfit import AbsaModel
82
+
83
+ # Download from the 🤗 Hub
84
+ model = AbsaModel.from_pretrained(
85
+ "Davide1999/setfit-absa-paraphrase-mpnet-base-v2-restaurants-aspect",
86
+ "Davide1999/setfit-absa-paraphrase-mpnet-base-v2-restaurants-polarity",
87
+ )
88
+ # Run inference
89
+ preds = model("The food was great, but the venue is just way too busy.")
90
+ ```
91
+
92
+ <!--
93
+ ### Downstream Use
94
+
95
+ *List how someone could finetune this model on their own dataset.*
96
+ -->
97
+
98
+ <!--
99
+ ### Out-of-Scope Use
100
+
101
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
102
+ -->
103
+
104
+ <!--
105
+ ## Bias, Risks and Limitations
106
+
107
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
108
+ -->
109
+
110
+ <!--
111
+ ### Recommendations
112
+
113
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
114
+ -->
115
+
116
+ ## Training Details
117
+
118
+ ### Training Set Metrics
119
+ | Training set | Min | Median | Max |
120
+ |:-------------|:----|:--------|:----|
121
+ | Word count | 4 | 17.9296 | 37 |
122
+
123
+ | Label | Training Sample Count |
124
+ |:----------|:----------------------|
125
+ | no aspect | 71 |
126
+ | aspect | 128 |
127
+
128
+ ### Training Hyperparameters
129
+ - batch_size: (16, 2)
130
+ - num_epochs: (1, 16)
131
+ - max_steps: -1
132
+ - sampling_strategy: oversampling
133
+ - body_learning_rate: (2e-05, 1e-05)
134
+ - head_learning_rate: 0.01
135
+ - loss: CosineSimilarityLoss
136
+ - distance_metric: cosine_distance
137
+ - margin: 0.25
138
+ - end_to_end: False
139
+ - use_amp: False
140
+ - warmup_proportion: 0.1
141
+ - seed: 42
142
+ - eval_max_steps: -1
143
+ - load_best_model_at_end: False
144
+
145
+ ### Training Results
146
+ | Epoch | Step | Training Loss | Validation Loss |
147
+ |:------:|:----:|:-------------:|:---------------:|
148
+ | 0.0007 | 1 | 0.3344 | - |
149
+ | 0.0370 | 50 | 0.2609 | - |
150
+ | 0.0740 | 100 | 0.1838 | - |
151
+ | 0.1109 | 150 | 0.0581 | - |
152
+ | 0.1479 | 200 | 0.0028 | - |
153
+ | 0.1849 | 250 | 0.0005 | - |
154
+ | 0.2219 | 300 | 0.0005 | - |
155
+ | 0.2589 | 350 | 0.0005 | - |
156
+ | 0.2959 | 400 | 0.0002 | - |
157
+ | 0.3328 | 450 | 0.0003 | - |
158
+ | 0.3698 | 500 | 0.0001 | - |
159
+ | 0.4068 | 550 | 0.0001 | - |
160
+ | 0.4438 | 600 | 0.0001 | - |
161
+ | 0.4808 | 650 | 0.0001 | - |
162
+ | 0.5178 | 700 | 0.0001 | - |
163
+ | 0.5547 | 750 | 0.0001 | - |
164
+ | 0.5917 | 800 | 0.0001 | - |
165
+ | 0.6287 | 850 | 0.0002 | - |
166
+ | 0.6657 | 900 | 0.0001 | - |
167
+ | 0.7027 | 950 | 0.0001 | - |
168
+ | 0.7396 | 1000 | 0.0001 | - |
169
+ | 0.7766 | 1050 | 0.0 | - |
170
+ | 0.8136 | 1100 | 0.0001 | - |
171
+ | 0.8506 | 1150 | 0.0001 | - |
172
+ | 0.8876 | 1200 | 0.0001 | - |
173
+ | 0.9246 | 1250 | 0.0001 | - |
174
+ | 0.9615 | 1300 | 0.0001 | - |
175
+ | 0.9985 | 1350 | 0.0001 | - |
176
+
177
+ ### Framework Versions
178
+ - Python: 3.10.12
179
+ - SetFit: 1.0.3
180
+ - Sentence Transformers: 3.0.1
181
+ - spaCy: 3.7.5
182
+ - Transformers: 4.39.0
183
+ - PyTorch: 2.3.0+cu121
184
+ - Datasets: 2.20.0
185
+ - Tokenizers: 0.15.2
186
+
187
+ ## Citation
188
+
189
+ ### BibTeX
190
+ ```bibtex
191
+ @article{https://doi.org/10.48550/arxiv.2209.11055,
192
+ doi = {10.48550/ARXIV.2209.11055},
193
+ url = {https://arxiv.org/abs/2209.11055},
194
+ author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
195
+ keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
196
+ title = {Efficient Few-Shot Learning Without Prompts},
197
+ publisher = {arXiv},
198
+ year = {2022},
199
+ copyright = {Creative Commons Attribution 4.0 International}
200
+ }
201
+ ```
202
+
203
+ <!--
204
+ ## Glossary
205
+
206
+ *Clearly define terms in order to be accessible across audiences.*
207
+ -->
208
+
209
+ <!--
210
+ ## Model Card Authors
211
+
212
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
213
+ -->
214
+
215
+ <!--
216
+ ## Model Card Contact
217
+
218
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
219
+ -->
config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "sentence-transformers/paraphrase-mpnet-base-v2",
3
+ "architectures": [
4
+ "MPNetModel"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-05,
15
+ "max_position_embeddings": 514,
16
+ "model_type": "mpnet",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 1,
20
+ "relative_attention_num_buckets": 32,
21
+ "torch_dtype": "float32",
22
+ "transformers_version": "4.39.0",
23
+ "vocab_size": 30527
24
+ }
config_sentence_transformers.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "3.0.1",
4
+ "transformers": "4.39.0",
5
+ "pytorch": "2.3.0+cu121"
6
+ },
7
+ "prompts": {},
8
+ "default_prompt_name": null,
9
+ "similarity_fn_name": null
10
+ }
config_setfit.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "span_context": 0,
3
+ "labels": [
4
+ "no aspect",
5
+ "aspect"
6
+ ],
7
+ "normalize_embeddings": false,
8
+ "spacy_model": "en_core_web_lg"
9
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af1aff0cb8d819c7fa3951b367b794625dd98e91d77b94a6bd9ba83cbbbad97d
3
+ size 437967672
model_head.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49ae4b4d33508540c8161404926d387e9a182d6c7cfb86e0ec6ae0264a65eef3
3
+ size 6991
modules.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ }
14
+ ]
sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 512,
3
+ "do_lower_case": false
4
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "[UNK]",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "104": {
28
+ "content": "[UNK]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "30526": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": true,
46
+ "cls_token": "<s>",
47
+ "do_basic_tokenize": true,
48
+ "do_lower_case": true,
49
+ "eos_token": "</s>",
50
+ "mask_token": "<mask>",
51
+ "model_max_length": 512,
52
+ "never_split": null,
53
+ "pad_token": "<pad>",
54
+ "sep_token": "</s>",
55
+ "strip_accents": null,
56
+ "tokenize_chinese_chars": true,
57
+ "tokenizer_class": "MPNetTokenizer",
58
+ "unk_token": "[UNK]"
59
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff