Spaces:
Sleeping
Sleeping
maximuspowers
commited on
Commit
•
090b1b7
1
Parent(s):
60ec004
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import nlpaug.augmenter.word as naw
|
3 |
+
|
4 |
+
# augmentations
|
5 |
+
def augment_text(text, aug_type, model_type=None, model_path=None, aug_p=0.25, aug_max=3):
|
6 |
+
if aug_type == 'Word Embedding Substitution':
|
7 |
+
aug = naw.WordEmbsAug(
|
8 |
+
model_type=model_type,
|
9 |
+
model_path=model_path,
|
10 |
+
action="substitute",
|
11 |
+
aug_p=aug_p
|
12 |
+
)
|
13 |
+
elif aug_type == 'Contextual Insertion':
|
14 |
+
aug = naw.ContextualWordEmbsAug(
|
15 |
+
model_path='bert-base-uncased',
|
16 |
+
action="insert",
|
17 |
+
aug_p=aug_p
|
18 |
+
)
|
19 |
+
elif aug_type == 'Synonym Replacement':
|
20 |
+
aug = naw.SynonymAug(
|
21 |
+
aug_src="wordnet",
|
22 |
+
aug_max=aug_max
|
23 |
+
)
|
24 |
+
elif aug_type == 'Back Translation':
|
25 |
+
aug = naw.BackTranslationAug(
|
26 |
+
from_model_name='facebook/wmt19-en-de',
|
27 |
+
to_model_name='facebook/wmt19-de-en'
|
28 |
+
)
|
29 |
+
else:
|
30 |
+
return text
|
31 |
+
|
32 |
+
augmented_text = aug.augment(text)
|
33 |
+
return augmented_text
|
34 |
+
|
35 |
+
with gr.Blocks() as iface:
|
36 |
+
text_input = gr.Textbox(label="Input Text")
|
37 |
+
aug_type_input = gr.Radio(
|
38 |
+
choices=['Word Embedding Substitution', 'Contextual Insertion', 'Synonym Replacement', 'Back Translation'],
|
39 |
+
label="Augmentation Type",
|
40 |
+
value='Word Embedding Substitution'
|
41 |
+
)
|
42 |
+
|
43 |
+
model_type_input = gr.Dropdown(
|
44 |
+
choices=['word2vec', 'fasttext', 'glove'],
|
45 |
+
label="Model Type (for Word Embedding Substitution)",
|
46 |
+
value='word2vec',
|
47 |
+
visible=True
|
48 |
+
)
|
49 |
+
model_path_input = gr.Textbox(
|
50 |
+
label="Model Path (for Word Embedding Substitution)",
|
51 |
+
value="GoogleNews-vectors-negative300.bin",
|
52 |
+
visible=True
|
53 |
+
)
|
54 |
+
aug_p_input = gr.Slider(
|
55 |
+
minimum=0, maximum=1, step=0.05, value=0.25,
|
56 |
+
label="Probability of Augmentation (for Embedding Substitution or Contextual Insertion)"
|
57 |
+
)
|
58 |
+
aug_max_input = gr.Slider(
|
59 |
+
minimum=1, maximum=10, step=1, value=3,
|
60 |
+
label="Max Number of Words to Change (for Synonym Replacement)",
|
61 |
+
visible=False
|
62 |
+
)
|
63 |
+
|
64 |
+
augmented_output = gr.Textbox(label="Augmented Text")
|
65 |
+
|
66 |
+
# update input block visibility based on aug type
|
67 |
+
def update_inputs(aug_type):
|
68 |
+
if aug_type == 'Word Embedding Substitution':
|
69 |
+
return gr.update(visible=True), gr.update(visible=True), gr.update(visible=False)
|
70 |
+
elif aug_type == 'Contextual Insertion':
|
71 |
+
return gr.update(visible=False), gr.update(visible=False), gr.update(visible=False)
|
72 |
+
elif aug_type == 'Synonym Replacement':
|
73 |
+
return gr.update(visible=False), gr.update(visible=False), gr.update(visible=True)
|
74 |
+
elif aug_type == 'Back Translation':
|
75 |
+
return gr.update(visible=False), gr.update(visible=False), gr.update(visible=False)
|
76 |
+
|
77 |
+
# update inputs when aug type changes
|
78 |
+
aug_type_input.change(
|
79 |
+
update_inputs,
|
80 |
+
inputs=[aug_type_input],
|
81 |
+
outputs=[model_type_input, model_path_input, aug_max_input]
|
82 |
+
)
|
83 |
+
|
84 |
+
apply_button = gr.Button("Apply Augmentation")
|
85 |
+
|
86 |
+
apply_button.click(
|
87 |
+
augment_text,
|
88 |
+
inputs=[text_input, aug_type_input, model_type_input, model_path_input, aug_p_input, aug_max_input],
|
89 |
+
outputs=[augmented_output]
|
90 |
+
)
|
91 |
+
|
92 |
+
iface.launch()
|