52 Commits

Author SHA1 Message Date
Stefan Schweter
2cfcba875e readme: finalize release of new ConvBERT model for Turkish - ConvBERTurk 2021-03-16 00:31:51 +01:00
Stefan Schweter
6ec96283ab Merge branch 'convbert-release' 2021-03-16 00:30:47 +01:00
Stefan Schweter
3bc709a751 readme: add asd significance tests for IMST test dataset 2021-03-15 23:45:45 +01:00
Stefan Schweter
373503bf61 figure: asd test for IMST dataset 2021-03-15 23:39:27 +01:00
Stefan Schweter
bf761069f1 readme: public release of new ConvBERTurk with new evaluations on downstream tasks 2021-03-13 02:09:57 +01:00
Stefan Schweter
d77757ab87 figure: add nice plots from evaluation on various downstream tasks 2021-03-13 02:05:48 +01:00
Stefan Schweter
75cef74522 script: add FLERT based training script 2021-03-13 00:54:34 +01:00
Stefan Schweter
3cc154da3c Merge branch 'electra' 2020-05-12 17:26:14 +02:00
Stefan Schweter
f1c27e09fb readme: add links to Hugging Face model hub 2020-05-12 17:05:23 +02:00
Stefan Schweter
a7921401ab readme: fix introduction section 2020-05-12 17:02:52 +02:00
Stefan Schweter
d00a0c4ecf electra: add configurations, used for ELECTRA training (small and base) 2020-05-12 16:56:07 +02:00
Stefan Schweter
1144b50942 electra: start with solid cheatsheet for pre-processing/pre-training ELECTRA models 2020-05-12 16:55:30 +02:00
Stefan Schweter
f258cf210d electra: slim down evaluation section 2020-05-12 16:20:12 +02:00
Stefan Schweter
2f1795356d readme: mention hyper-parameters used for fine-tuning models 2020-05-12 16:19:40 +02:00
Stefan Schweter
e897374956 electra: minor fixes 2020-05-12 16:10:56 +02:00
Stefan Schweter
58d2f10ea9 readme: add new model usage section 2020-05-12 16:10:41 +02:00
Stefan Schweter
49bcc08443 configs: delete outdated configs 2020-05-12 12:22:30 +02:00
Stefan Schweter
7ebfda4f56 readme: mention ELECTRA release. Update model comparison section with 🤗/Transformers evaluation 2020-05-12 12:21:39 +02:00
Stefan Schweter
0d5fd2159e electra: fix image alt tag. Remove model comparison section 2020-05-12 12:20:31 +02:00
Stefan Schweter
f209d4b2c4 electra: add final results for PoS tagging 2020-05-09 14:27:24 +02:00
Stefan Schweter
569820b4a9 electra: add results on downstream tasks. Add model comparison section and model loss section 2020-05-09 00:56:21 +02:00
Stefan Schweter
4741c695b1 figures: add plots for performance of small/base models on downstream tasks. Also add loss curves for small/base models 2020-05-09 00:55:29 +02:00
Stefan Schweter
8cac4dc513 electra: saner presentation of results ;) 2020-05-08 01:18:45 +02:00
Stefan Schweter
055b7e5183 electra: adds results for best checkpoints (PoS tagging and NER) 2020-05-08 01:13:41 +02:00
Stefan Schweter
c8fb7e1a43 electra: minor markdown fix 2020-05-08 01:02:50 +02:00
Stefan Schweter
9bd0586d0b figures: update plots 2020-05-08 01:02:28 +02:00
Stefan Schweter
67481b43fe electra: clarify evaluation over 5 different runs with averaged metrics 2020-05-08 00:57:34 +02:00
Stefan Schweter
a2cdb5f087 electra: fix link to TensorBoard for ELECTRA small model 2020-05-08 00:36:26 +02:00
Stefan Schweter
95e38c0cb8 electra: add initial version of readme (incl. ELECTRA small model) 2020-05-07 23:52:45 +02:00
Stefan Schweter
c7c0e97016 figures: add plots for ELECTRA small on PoS tagging and NER tasks 2020-05-07 23:52:10 +02:00
Stefan Schweter
ec378ccaab readme: fix link to cheatsheet 2020-04-28 00:18:41 +02:00
Stefan Schweter
41e35a5190 readme: add citation section with doi 2020-04-27 23:19:30 +02:00
Stefan Schweter
302bf9eb50 readme: add reference to Zenodo 2020-04-27 23:06:50 +02:00
Stefan Schweter
bfb72ceffb readme: add link to cheetsheat 1.0.0 2020-04-27 22:57:33 +02:00
Stefan Schweter
9903e1b5dd cheatsheet: fix output name for uncased vocab 2020-04-27 22:56:00 +02:00
Stefan Schweter
fe6a8c3fbf readme: new release of uncased BERTurk model and BERTurk models with larger vocab size (128k, cased and uncased) 2020-03-25 14:20:00 +01:00
Stefan Schweter
ad7458e261 Merge pull request #9 from stefan-it/distilberturk-release
DistilBERTurk release
2020-03-11 23:06:48 +01:00
Stefan Schweter
fbed221906 readme: fix release date of DistilBERTurk model 2020-03-11 23:03:21 +01:00
Stefan Schweter
216397d5e8 configs: add FARM configurations for DistilBERTurk 2020-03-11 23:02:17 +01:00
Stefan Schweter
243ed29742 readme: new release of (cased) distilled BERTurk model: DistilBERTurk 2020-03-10 18:38:09 +01:00
Stefan Schweter
7e03b7e585 readme: minor fix 2020-02-17 01:17:23 +01:00
Stefan Schweter
445907d775 cheatsheet: properly mention sentence splitting with NLTK and removal of short sentences 2020-02-17 00:28:53 +01:00
Stefan Schweter
16155079ac configs: add FARM configurations for PoS tagging task 2020-02-17 00:25:59 +01:00
Stefan Schweter
a364f76566 readme: update results for PoS tagging and NER. Prepare for base model release of BERTurk, a community-driven model from the awesome Turkish NLP comminity 2020-02-17 00:12:37 +01:00
Stefan Schweter
a11daf138d configs: add FARM configurations for PoS tagging task 2020-02-17 00:11:06 +01:00
Stefan Schweter
148a9c164d readme: february update (new models are training) 2020-02-10 15:26:31 +01:00
Stefan Schweter
dc03a46322 readme: add steps for uncased model (pretraining and training) 2020-02-06 10:45:27 +01:00
Stefan Schweter
f7d1a49c53 readme: add new sections for BERT pretraining (TPU + VM creation, start pretraining on TPU) 2020-02-05 19:40:34 +01:00
Stefan Schweter
84325c975f readme: add preprocessing section to cheatsheet 2020-02-04 17:26:16 +01:00
Stefan Schweter
d38bc271ff readme: add cheatsheet for upcoming model 2020-02-04 14:42:07 +01:00