RagN commited on
Commit
f37472a
1 Parent(s): 133e047

Add SetFit ABSA model

Browse files
1_Pooling/config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
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
+ }
README.md ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: setfit
3
+ tags:
4
+ - setfit
5
+ - absa
6
+ - sentence-transformers
7
+ - text-classification
8
+ - generated_from_setfit_trainer
9
+ metrics:
10
+ - accuracy
11
+ widget:
12
+ - text: is the best French food you will find:It may be a bit packed on weekends,
13
+ but the vibe is good and it is the best French food you will find in the area.
14
+ - text: knew what the specials were.:Whem asked, we had to ask more detailed questions
15
+ so that we knew what the specials were.
16
+ - text: all out wow dining experience.:Go here for a romantic dinner but not for an
17
+ all out wow dining experience.
18
+ - text: vibe, the owner is super friendly:Best of all is the warm vibe, the owner
19
+ is super friendly and service is fast.
20
+ - text: all of the dishes are excellent.:The menu is limited but almost all of the
21
+ dishes are excellent.
22
+ pipeline_tag: text-classification
23
+ inference: false
24
+ base_model: sentence-transformers/paraphrase-mpnet-base-v2
25
+ ---
26
+
27
+ # SetFit Polarity Model with sentence-transformers/paraphrase-mpnet-base-v2
28
+
29
+ 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 classifying aspect polarities.
30
+
31
+ The model has been trained using an efficient few-shot learning technique that involves:
32
+
33
+ 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
34
+ 2. Training a classification head with features from the fine-tuned Sentence Transformer.
35
+
36
+ This model was trained within the context of a larger system for ABSA, which looks like so:
37
+
38
+ 1. Use a spaCy model to select possible aspect span candidates.
39
+ 2. Use a SetFit model to filter these possible aspect span candidates.
40
+ 3. **Use this SetFit model to classify the filtered aspect span candidates.**
41
+
42
+ ## Model Details
43
+
44
+ ### Model Description
45
+ - **Model Type:** SetFit
46
+ - **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2)
47
+ - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
48
+ - **spaCy Model:** en_core_web_lg
49
+ - **SetFitABSA Aspect Model:** [/var/folders/df/ptj1_nrd3cvbsnwm3j_0dtv00000gq/T/tmp038p09w0/RagN/ABSA-aspect](https://huggingface.co//var/folders/df/ptj1_nrd3cvbsnwm3j_0dtv00000gq/T/tmp038p09w0/RagN/ABSA-aspect)
50
+ - **SetFitABSA Polarity Model:** [/var/folders/df/ptj1_nrd3cvbsnwm3j_0dtv00000gq/T/tmp038p09w0/RagN/ABSA-polarity](https://huggingface.co//var/folders/df/ptj1_nrd3cvbsnwm3j_0dtv00000gq/T/tmp038p09w0/RagN/ABSA-polarity)
51
+ - **Maximum Sequence Length:** 512 tokens
52
+ - **Number of Classes:** 4 classes
53
+ <!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
54
+ <!-- - **Language:** Unknown -->
55
+ <!-- - **License:** Unknown -->
56
+
57
+ ### Model Sources
58
+
59
+ - **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
60
+ - **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
61
+ - **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
62
+
63
+ ### Model Labels
64
+ | Label | Examples |
65
+ |:---------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
66
+ | negative | <ul><li>'But the staff was so horrible:But the staff was so horrible to us.'</li><li>', forgot our toast, left out:They did not have mayonnaise, forgot our toast, left out ingredients (ie cheese in an omelet), below hot temperatures and the bacon was so over cooked it crumbled on the plate when you touched it.'</li><li>'did not have mayonnaise, forgot our:They did not have mayonnaise, forgot our toast, left out ingredients (ie cheese in an omelet), below hot temperatures and the bacon was so over cooked it crumbled on the plate when you touched it.'</li></ul> |
67
+ | positive | <ul><li>"factor was the food, which was: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>"The food is uniformly exceptional: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><li>"a very capable kitchen which will proudly: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> |
68
+ | neutral | <ul><li>"'s on the menu or not.: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><li>'to sample both meats).:Our agreed favorite is the orrechiete with sausage and chicken (usually the waiters are kind enough to split the dish in half so you get to sample both meats).'</li><li>'to split the dish in half so:Our agreed favorite is the orrechiete with sausage and chicken (usually the waiters are kind enough to split the dish in half so you get to sample both meats).'</li></ul> |
69
+ | conflict | <ul><li>'The food was delicious but:The food was delicious but do not come here on a empty stomach.'</li><li>"The service varys from day:The service varys from day to day- sometimes they're very nice, and sometimes not."</li></ul> |
70
+
71
+ ## Uses
72
+
73
+ ### Direct Use for Inference
74
+
75
+ First install the SetFit library:
76
+
77
+ ```bash
78
+ pip install setfit
79
+ ```
80
+
81
+ Then you can load this model and run inference.
82
+
83
+ ```python
84
+ from setfit import AbsaModel
85
+
86
+ # Download from the 🤗 Hub
87
+ model = AbsaModel.from_pretrained(
88
+ "/var/folders/df/ptj1_nrd3cvbsnwm3j_0dtv00000gq/T/tmp038p09w0/RagN/ABSA-aspect",
89
+ "/var/folders/df/ptj1_nrd3cvbsnwm3j_0dtv00000gq/T/tmp038p09w0/RagN/ABSA-polarity",
90
+ )
91
+ # Run inference
92
+ preds = model("The food was great, but the venue is just way too busy.")
93
+ ```
94
+
95
+ <!--
96
+ ### Downstream Use
97
+
98
+ *List how someone could finetune this model on their own dataset.*
99
+ -->
100
+
101
+ <!--
102
+ ### Out-of-Scope Use
103
+
104
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
105
+ -->
106
+
107
+ <!--
108
+ ## Bias, Risks and Limitations
109
+
110
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
111
+ -->
112
+
113
+ <!--
114
+ ### Recommendations
115
+
116
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
117
+ -->
118
+
119
+ ## Training Details
120
+
121
+ ### Training Set Metrics
122
+ | Training set | Min | Median | Max |
123
+ |:-------------|:----|:--------|:----|
124
+ | Word count | 6 | 21.3594 | 43 |
125
+
126
+ | Label | Training Sample Count |
127
+ |:---------|:----------------------|
128
+ | conflict | 2 |
129
+ | negative | 19 |
130
+ | neutral | 25 |
131
+ | positive | 82 |
132
+
133
+ ### Training Hyperparameters
134
+ - batch_size: (16, 2)
135
+ - num_epochs: (1, 16)
136
+ - max_steps: -1
137
+ - sampling_strategy: oversampling
138
+ - body_learning_rate: (2e-05, 1e-05)
139
+ - head_learning_rate: 0.01
140
+ - loss: CosineSimilarityLoss
141
+ - distance_metric: cosine_distance
142
+ - margin: 0.25
143
+ - end_to_end: False
144
+ - use_amp: False
145
+ - warmup_proportion: 0.1
146
+ - seed: 42
147
+ - eval_max_steps: -1
148
+ - load_best_model_at_end: False
149
+
150
+ ### Training Results
151
+ | Epoch | Step | Training Loss | Validation Loss |
152
+ |:------:|:----:|:-------------:|:---------------:|
153
+ | 0.0018 | 1 | 0.2117 | - |
154
+ | 0.0923 | 50 | 0.1022 | - |
155
+ | 0.1845 | 100 | 0.0602 | - |
156
+ | 0.2768 | 150 | 0.0012 | - |
157
+ | 0.3690 | 200 | 0.0003 | - |
158
+ | 0.4613 | 250 | 0.0004 | - |
159
+ | 0.5535 | 300 | 0.0007 | - |
160
+ | 0.6458 | 350 | 0.0004 | - |
161
+ | 0.7380 | 400 | 0.0005 | - |
162
+ | 0.8303 | 450 | 0.0002 | - |
163
+ | 0.9225 | 500 | 0.0002 | - |
164
+
165
+ ### Framework Versions
166
+ - Python: 3.9.6
167
+ - SetFit: 1.0.3
168
+ - Sentence Transformers: 2.3.1
169
+ - spaCy: 3.7.2
170
+ - Transformers: 4.36.2
171
+ - PyTorch: 2.1.1
172
+ - Datasets: 2.16.1
173
+ - Tokenizers: 0.15.0
174
+
175
+ ## Citation
176
+
177
+ ### BibTeX
178
+ ```bibtex
179
+ @article{https://doi.org/10.48550/arxiv.2209.11055,
180
+ doi = {10.48550/ARXIV.2209.11055},
181
+ url = {https://arxiv.org/abs/2209.11055},
182
+ author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
183
+ keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
184
+ title = {Efficient Few-Shot Learning Without Prompts},
185
+ publisher = {arXiv},
186
+ year = {2022},
187
+ copyright = {Creative Commons Attribution 4.0 International}
188
+ }
189
+ ```
190
+
191
+ <!--
192
+ ## Glossary
193
+
194
+ *Clearly define terms in order to be accessible across audiences.*
195
+ -->
196
+
197
+ <!--
198
+ ## Model Card Authors
199
+
200
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
201
+ -->
202
+
203
+ <!--
204
+ ## Model Card Contact
205
+
206
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
207
+ -->
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": "4.36.2",
23
+ "vocab_size": 30527
24
+ }
config_sentence_transformers.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "2.0.0",
4
+ "transformers": "4.7.0",
5
+ "pytorch": "1.9.0+cu102"
6
+ }
7
+ }
config_setfit.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "normalize_embeddings": false,
3
+ "span_context": 3,
4
+ "spacy_model": "en_core_web_lg",
5
+ "labels": [
6
+ "conflict",
7
+ "negative",
8
+ "neutral",
9
+ "positive"
10
+ ]
11
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12309202370790e63aea91c0f5a01da2772ffae4ff2dc1619caf48d6bd26acc6
3
+ size 437967672
model_head.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a964fac7c68e33b2e72e37b20d48cdaff0888408db911ad90083d15999755f38
3
+ size 25559
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