alihan/textattack-nlp-transformer
1
0
Fork 0
mirror of https://github.com/QData/TextAttack.git synced 2021-10-13 00:05:06 +03:00
Code Issues Projects Releases Wiki Activity
1,082 Commits 20 Branches 13 Tags
46b0f389672dd7f7df5b360061f497e16c4e04c0
Commit Graph

34 Commits

Author SHA1 Message Date
Jack Morris
462620e074 trimming the fat; removing excess stuff; changing names 2020-06-18 11:12:36 -04:00
Jack Morris
90e6872eb4 ugly merge with recent master 2020-06-17 21:45:50 -04:00
Jack Morris
d6e90065c2 format pre-merge 2020-06-17 19:47:33 -04:00
Jack Morris
07e927e3e4 add isort and fix import loops, preparing to shorten tests 2020-06-15 20:40:44 -04:00
Jack Morris
1d8a172c05 code formatting; tokenizers->models/tokenizers 2020-06-15 18:36:04 -04:00
Jack Morris
fcb82dac5e makefile and setup; need to fix imports 2020-06-15 17:07:12 -04:00
Jack Morris
c4329985a0 working on SST models and parallel memory issue 2020-06-15 13:55:07 -04:00
Jack Morris
889954aaa9 test huggingface models; test attack from file; remove old files 2020-06-10 15:34:52 -04:00
Jack Morris
8b924362a5 standardize tokenization 2020-06-10 10:16:11 -04:00
Jack Morris
f4d7ba0f1a add --attack-from-file, unify tokenizers, add models from huggingface model hub [still need to verify acc] 2020-06-07 15:12:02 -04:00
Jack Morris
62d3a1070c README; get_logger()->logger 2020-06-05 20:12:14 -04:00
Jack Morris
6d225c3345 add tokenizers & models for text2text, init goal function & validators 2020-04-26 18:53:59 -04:00
Jack Morris
c661180d30 undo other changes 2020-03-24 17:44:23 -04:00
Jack Morris
f2ebf5e284 cache constraint results 2020-03-23 13:47:37 -04:00
Jin Yong (Jeffrey) Yoo
7dce0808e4 Disable fast tokenizer 
Because fast tokenizer cannot be pickled, it cannot be used with multiprocessing.
 2020-03-16 20:22:01 -04:00
Jin Yong Yoo
1171e1270a introduce faster BERT tokenizer 2020-03-14 23:44:07 -04:00
uvafan
0b6da2a475 fix model 2020-03-05 22:29:07 -05:00
Jack Morris
e5cda5731b remove cf synonym sampling, log hparams to tb, train on full set of paraphrases 2020-03-05 09:34:04 -05:00
Jack Morris
d82cb57d6d make recipes work, move config into root folder 2020-01-19 15:20:59 -05:00
Jack Morris
ad75e69b1b reorganize: scripts/ folder; shared/ folder; attack_methods/ instead of attack/; only take a single transformation as an argument 2020-01-19 14:25:30 -05:00
Jack Morris
3c8180a293 change printable() to clean_text(), fix typo in comment, change SPLIT_TOKEN 2020-01-13 11:36:37 -05:00
Jack Morris
0d3fd16073 remove print statements 2020-01-10 09:29:39 -05:00
Jack Morris
6c9283d729 finish batching for entailment 2020-01-10 09:09:20 -05:00
Jack Morris
205ec7aeaa benchmark entailment; upload datasets and models for MNLI and SNLI 2020-01-09 16:58:24 -05:00
Jack Morris
1de5194f64 smarter way to manage tokenizedtext; correct labels for entailment 2020-01-09 16:06:40 -05:00
Jack Morris
633f56d2e5 merge in newest stuff 2020-01-09 13:23:15 -05:00
Jack Morris
da1613d663 upload word vectors, CNN, LSTM 2020-01-08 18:28:54 -05:00
Jack Morris
dffdbecbdd initial go at entailment: add MNLI 2019-12-02 23:32:07 -05:00
Jack Morris
672cab5bee add AG News dataset 2019-12-02 08:54:30 -05:00
Jack Morris
a2e3ec739c move tokenizer out of model class and fix some bugs 2019-11-29 16:53:00 -05:00
Jack Morris
88996d4a05 separate model and tokenizer functionality 2019-11-29 16:02:52 -05:00
Jack Morris
095f5b4d11 change max_seq_length to 256 2019-11-19 17:30:28 -05:00
Jack Morris
6a5bbb27a4 add good bert models and softmax 2019-11-19 00:22:49 -05:00
Jack Morris
17305982b0 generic bert class and bert for MR; also a script for testing models 2019-11-08 17:38:54 -05:00