Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • 0cde083c4216bb065e67b2da3be4ebd30bbe60b3
  • master default protected
  • add_default_vectorizer_and_pretrained_loading
  • change_system_act_in_env
  • clean_code
  • convlab_exp
  • crossowoz_ddpt
  • e2e-soloist
  • emoUS
  • generalized_action_dicts
  • issue127
  • issue_114
  • nlg-scgpt
  • ppo_num_dialogues
  • pre-training
  • readme
  • remapping_actions
  • robust_masking_feature
  • scgpt_exp
  • soloist
20 results
Created with Raphaël 2.2.019Feb1817161514102130Jan2624211913731Dec2827222122212220191514109872130Nov2928272413Sep19Aug30Jul718Jun1720May729Apr16215Mar1428Jan29Dec18Nov131Oct30292827262218171513762130Sep2928272524232221201910987542129Aug26241916629Jul27262422181615131097429Jun23181715131025May144229Apr129FebMerge branch 'master' into unified_datasetMerge pull request #28 from ConvLab/remap_book_actionsbooking remapper checks if act exists in ontologybufix in booking_remapper. And i can not remap 134 actions in the data because of wrong labellingfirst version of booking_remapper, there are some assertion errors due to incorrect labelling of dataupdate unified data: remove value in binary dialog acts. update ontology: use boolean 'user' and 'system' keys instead of 'speaker' in dialog actsfirst version that can convert multiwoz data, trains supervised model and evaluates model with simulated user, f1-score is 0.52 and success rate is 73%remapping_actionsremapping_actionsworking MLE version with remapped data set, getting F1-score of 0.52Merge pull request #26 from ConvLab/unified_datasetcan set an occurance bound when created the act setupdate sgd, tm1/2/3 ontologyadded remapping of span actsrefactored remap_actionsfirst version of action remapping in multiwoz, will now refactor it and then create the system and user acts from the remapped dataMerge pull request #25 from ConvLab/unified_datasetadd dialog act dictionary in ontologyUpdate multiwoz_eval.Merge branch 'github_master' into 'master'Update evaluate_distributed.py, needed to set torch.seed.Merge pull request #24 from ConvLab/complete_rate_refactorcomplete rate is now also evaluated by multiwoz_evaluator and not user to make it more consistent and easier to readUpdate evaluate_distributed.py, needed to set torch.seed.Merge pull request #23 from ConvLab/unified_datasetupdate multiwoz21 preprocessadd booked information for multiwoz21add tm3, update tm1&tm2add tm2 datasetupdate tm1 datasetupdate readme for repo linkMerge pull request #22 from ConvLab/dsml_convlabadd taskmaster-1(tm1)Merge branch 'master' into dsml_convlabMerge pull request #1 from ConvLab/dsml_scgptAdd DSML Group Codetrain bert for bio tagging. add utf-8 for dumping jsontrain bert for token classificationadd package version checkclean up run_seq2seq.pyupdate run_rg.sh: add create_data scriptbase model run response generation