Skip to content
This repository has been archived by the owner on Dec 24, 2023. It is now read-only.

Latest commit

 

History

History
363 lines (322 loc) · 13.4 KB

README.md

File metadata and controls

363 lines (322 loc) · 13.4 KB

TextAttack Model Zoo

More details at https://textattack.readthedocs.io/en/latest/3recipes/models.html

TextAttack includes pre-trained models for different common NLP tasks. This makes it easier for users to get started with TextAttack. It also enables a more fair comparison of attacks from the literature.

All evaluation results were obtained using textattack eval to evaluate models on their default test dataset (test set, if labels are available, otherwise, eval/validation set). You can use this command to verify the accuracies for yourself: for example, textattack eval --model roberta-base-mr.

The LSTM and wordCNN models' code is available in textattack.models.helpers. All other models are transformers imported from the transformers package. To list evaluate all TextAttack pretrained models, invoke textattack eval without specifying a model: textattack eval --num-examples 1000. All evaluations shown are on the full validation or test set up to 1000 examples.

LSTM

  • AG News (lstm-ag-news)
    • datasets dataset ag_news, split test
      • True Positive/Positive: 914/1000
      • Accuracy: 91.4%
  • IMDB (lstm-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 883/1000
      • Accuracy: 88.30%
  • Movie Reviews [Rotten Tomatoes] (lstm-mr)
    • datasets dataset rotten_tomatoes, split validation
      • True Positive/Positive: 807/1000
      • Accuracy: 80.70%
    • datasets dataset rotten_tomatoes, split test
      • True Positive/Positive: 781/1000
      • Accuracy: 78.10%
  • SST-2 (lstm-sst2)
    • datasets dataset glue, subset sst2, split validation
      • True Positive/Positive: 737/872
      • Accuracy: 84.52%
  • Yelp Polarity (lstm-yelp)
    • datasets dataset yelp_polarity, split test
      • True Positive/Positive: 922/1000
      • Accuracy: 92.20%

wordCNN

  • AG News (cnn-ag-news)
    • datasets dataset ag_news, split test
      • True Positive/Positive: 910/1000
      • Accuracy: 91.00%
  • IMDB (cnn-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 863/1000
      • Accuracy: 86.30%
  • Movie Reviews [Rotten Tomatoes] (cnn-mr)
    • datasets dataset rotten_tomatoes, split validation
      • True Positive/Positive: 794/1000
      • Accuracy: 79.40%
    • datasets dataset rotten_tomatoes, split test
      • True Positive/Positive: 768/1000
      • Accuracy: 76.80%
  • SST-2 (cnn-sst2)
    • datasets dataset glue, subset sst2, split validation
      • True Positive/Positive: 721/872
      • Accuracy: 82.68%
  • Yelp Polarity (cnn-yelp)
    • datasets dataset yelp_polarity, split test
      • True Positive/Positive: 913/1000
      • Accuracy: 91.30%

albert-base-v2

  • AG News (albert-base-v2-ag-news)
    • datasets dataset ag_news, split test
      • True Positive/Positive: 943/1000
      • Accuracy: 94.30%
  • CoLA (albert-base-v2-cola)
    • datasets dataset glue, subset cola, split validation
      • True Positive/Positive: 829/1000
      • Accuracy: 82.90%
  • IMDB (albert-base-v2-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 913/1000
      • Accuracy: 91.30%
  • Movie Reviews [Rotten Tomatoes] (albert-base-v2-mr)
    • datasets dataset rotten_tomatoes, split validation
      • True Positive/Positive: 882/1000
      • Accuracy: 88.20%
    • datasets dataset rotten_tomatoes, split test
      • True Positive/Positive: 851/1000
      • Accuracy: 85.10%
  • Quora Question Pairs (albert-base-v2-qqp)
    • datasets dataset glue, subset qqp, split validation
      • True Positive/Positive: 914/1000
      • Accuracy: 91.40%
  • Recognizing Textual Entailment (albert-base-v2-rte)
    • datasets dataset glue, subset rte, split validation
      • True Positive/Positive: 211/277
      • Accuracy: 76.17%
  • SNLI (albert-base-v2-snli)
    • datasets dataset snli, split test
      • True Positive/Positive: 883/1000
      • Accuracy: 88.30%
  • SST-2 (albert-base-v2-sst2)
    • datasets dataset glue, subset sst2, split validation
      • True Positive/Positive: 807/872
      • Accuracy: 92.55%)
  • STS-b (albert-base-v2-stsb)
    • datasets dataset glue, subset stsb, split validation
    • Pearson correlation: 0.9041359738552746
    • Spearman correlation: 0.8995912861209745
  • WNLI (albert-base-v2-wnli)
    • datasets dataset glue, subset wnli, split validation
      • True Positive/Positive: 42/71
      • Accuracy: 59.15%
  • Yelp Polarity (albert-base-v2-yelp)
    • datasets dataset yelp_polarity, split test
      • True Positive/Positive: 963/1000
      • Accuracy: 96.30%

bert-base-uncased

  • AG News (bert-base-uncased-ag-news)
    • datasets dataset ag_news, split test
      • True Positive/Positive: 942/1000
      • Accuracy: 94.20%
  • CoLA (bert-base-uncased-cola)
    • datasets dataset glue, subset cola, split validation
      • True Positive/Positive: 812/1000
      • Accuracy: 81.20%
  • IMDB (bert-base-uncased-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 919/1000
      • Accuracy: 91.90%
  • MNLI matched (bert-base-uncased-mnli)
    • datasets dataset glue, subset mnli, split validation_matched
      • True Positive/Positive: 840/1000
      • Accuracy: 84.00%
  • Movie Reviews [Rotten Tomatoes] (bert-base-uncased-mr)
    • datasets dataset rotten_tomatoes, split validation
      • True Positive/Positive: 876/1000
      • Accuracy: 87.60%
    • datasets dataset rotten_tomatoes, split test
      • True Positive/Positive: 838/1000
      • Accuracy: 83.80%
  • MRPC (bert-base-uncased-mrpc)
    • datasets dataset glue, subset mrpc, split validation
      • True Positive/Positive: 358/408
      • Accuracy: 87.75%
  • QNLI (bert-base-uncased-qnli)
    • datasets dataset glue, subset qnli, split validation
      • True Positive/Positive: 904/1000
      • Accuracy: 90.40%
  • Quora Question Pairs (bert-base-uncased-qqp)
    • datasets dataset glue, subset qqp, split validation
      • True Positive/Positive: 924/1000
      • Accuracy: 92.40%
  • Recognizing Textual Entailment (bert-base-uncased-rte)
    • datasets dataset glue, subset rte, split validation
      • True Positive/Positive: 201/277
      • Accuracy: 72.56%
  • SNLI (bert-base-uncased-snli)
    • datasets dataset snli, split test
      • True Positive/Positive: 894/1000
      • Accuracy: 89.40%
  • SST-2 (bert-base-uncased-sst2)
    • datasets dataset glue, subset sst2, split validation
      • True Positive/Positive: 806/872
      • Accuracy: 92.43%)
  • STS-b (bert-base-uncased-stsb)
    • datasets dataset glue, subset stsb, split validation
    • Pearson correlation: 0.8775458937815515
    • Spearman correlation: 0.8773251339980935
  • WNLI (bert-base-uncased-wnli)
    • datasets dataset glue, subset wnli, split validation
      • True Positive/Positive: 40/71
      • Accuracy: 56.34%
  • Yelp Polarity (bert-base-uncased-yelp)
    • datasets dataset yelp_polarity, split test
      • True Positive/Positive: 963/1000
      • Accuracy: 96.30%

distilbert-base-cased

  • CoLA (distilbert-base-cased-cola)
    • datasets dataset glue, subset cola, split validation
      • True Positive/Positive: 786/1000
      • Accuracy: 78.60%
  • MRPC (distilbert-base-cased-mrpc)
    • datasets dataset glue, subset mrpc, split validation
      • True Positive/Positive: 320/408
      • Accuracy: 78.43%
  • Quora Question Pairs (distilbert-base-cased-qqp)
    • datasets dataset glue, subset qqp, split validation
      • True Positive/Positive: 908/1000
      • Accuracy: 90.80%
  • SNLI (distilbert-base-cased-snli)
    • datasets dataset snli, split test
      • True Positive/Positive: 861/1000
      • Accuracy: 86.10%
  • SST-2 (distilbert-base-cased-sst2)
    • datasets dataset glue, subset sst2, split validation
      • True Positive/Positive: 785/872
      • Accuracy: 90.02%)
  • STS-b (distilbert-base-cased-stsb)
    • datasets dataset glue, subset stsb, split validation
    • Pearson correlation: 0.8421540899520146
    • Spearman correlation: 0.8407155030382939

distilbert-base-uncased

  • AG News (distilbert-base-uncased-ag-news)
    • datasets dataset ag_news, split test
      • True Positive/Positive: 944/1000
      • Accuracy: 94.40%
  • CoLA (distilbert-base-uncased-cola)
    • datasets dataset glue, subset cola, split validation
      • True Positive/Positive: 786/1000
      • Accuracy: 78.60%
  • IMDB (distilbert-base-uncased-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 903/1000
      • Accuracy: 90.30%
  • MNLI matched (distilbert-base-uncased-mnli)
    • datasets dataset glue, subset mnli, split validation_matched
      • True Positive/Positive: 817/1000
      • Accuracy: 81.70%
  • MRPC (distilbert-base-uncased-mrpc)
    • datasets dataset glue, subset mrpc, split validation
      • True Positive/Positive: 350/408
      • Accuracy: 85.78%
  • QNLI (distilbert-base-uncased-qnli)
    • datasets dataset glue, subset qnli, split validation
      • True Positive/Positive: 860/1000
      • Accuracy: 86.00%
  • Recognizing Textual Entailment (distilbert-base-uncased-rte)
    • datasets dataset glue, subset rte, split validation
      • True Positive/Positive: 180/277
      • Accuracy: 64.98%
  • STS-b (distilbert-base-uncased-stsb)
    • datasets dataset glue, subset stsb, split validation
    • Pearson correlation: 0.8421540899520146
    • Spearman correlation: 0.8407155030382939
  • WNLI (distilbert-base-uncased-wnli)
    • datasets dataset glue, subset wnli, split validation
      • True Positive/Positive: 40/71
      • Accuracy: 56.34%

roberta-base

  • AG News (roberta-base-ag-news)
    • datasets dataset ag_news, split test
      • True Positive/Positive: 947/1000
      • Accuracy: 94.70%
  • CoLA (roberta-base-cola)
    • datasets dataset glue, subset cola, split validation
      • True Positive/Positive: 857/1000
      • Accuracy: 85.70%
  • IMDB (roberta-base-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 941/1000
      • Accuracy: 94.10%
  • Movie Reviews [Rotten Tomatoes] (roberta-base-mr)
    • datasets dataset rotten_tomatoes, split validation
      • True Positive/Positive: 899/1000
      • Accuracy: 89.90%
    • datasets dataset rotten_tomatoes, split test
      • True Positive/Positive: 883/1000
      • Accuracy: 88.30%
  • MRPC (roberta-base-mrpc)
    • datasets dataset glue, subset mrpc, split validation
      • True Positive/Positive: 371/408
      • Accuracy: 91.18%
  • QNLI (roberta-base-qnli)
    • datasets dataset glue, subset qnli, split validation
      • True Positive/Positive: 917/1000
      • Accuracy: 91.70%
  • Recognizing Textual Entailment (roberta-base-rte)
    • datasets dataset glue, subset rte, split validation
      • True Positive/Positive: 217/277
      • Accuracy: 78.34%
  • SST-2 (roberta-base-sst2)
    • datasets dataset glue, subset sst2, split validation
      • True Positive/Positive: 820/872
      • Accuracy: 94.04%)
  • STS-b (roberta-base-stsb)
    • datasets dataset glue, subset stsb, split validation
    • Pearson correlation: 0.906067852162708
    • Spearman correlation: 0.9025045272903051
  • WNLI (roberta-base-wnli)
    • datasets dataset glue, subset wnli, split validation
      • True Positive/Positive: 40/71
      • Accuracy: 56.34%

xlnet-base-cased

  • CoLA (xlnet-base-cased-cola)
    • datasets dataset glue, subset cola, split validation
      • True Positive/Positive: 800/1000
      • Accuracy: 80.00%
  • IMDB (xlnet-base-cased-imdb)
    • datasets dataset imdb, split test
      • True Positive/Positive: 957/1000
      • Accuracy: 95.70%
  • Movie Reviews [Rotten Tomatoes] (xlnet-base-cased-mr)
    • datasets dataset rotten_tomatoes, split validation
      • True Positive/Positive: 908/1000
      • Accuracy: 90.80%
    • datasets dataset rotten_tomatoes, split test
      • True Positive/Positive: 876/1000
      • Accuracy: 87.60%
  • MRPC (xlnet-base-cased-mrpc)
    • datasets dataset glue, subset mrpc, split validation
      • True Positive/Positive: 363/408
      • Accuracy: 88.97%
  • Recognizing Textual Entailment (xlnet-base-cased-rte)
    • datasets dataset glue, subset rte, split validation
      • True Positive/Positive: 196/277
      • Accuracy: 70.76%
  • STS-b (xlnet-base-cased-stsb)
    • datasets dataset glue, subset stsb, split validation
    • Pearson correlation: 0.883111673280641
    • Spearman correlation: 0.8773439961182335
  • WNLI (xlnet-base-cased-wnli)
    • datasets dataset glue, subset wnli, split validation
      • True Positive/Positive: 41/71
      • Accuracy: 57.75%