Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • master default protected
1 result
Created with Raphaël 2.2.031Mar3029282723201716141312109876428Feb262521181514Add project reportmastermasterAdd missing module documentationCleanup, PEP 8 conventions, commentsExplicitly mention NEC results in READMEAdd Word2Vec resultsAdd finetuning resultsMerge remote-tracking branch 'origin/master'New NEC and NER results after FIGER dataset fixLoading of finetuned models and minor fixesMerge remote-tracking branch 'origin/master'Fixed FIGER dataset problem where there are multiple possible labels for an entity and only one is counted as correctUpdate loss plot- Pushed logs from NEC test runs- Merged NEC testcases to oneMerge branch 'master' of gitlab.cl.uni-heidelberg.de:fsem25-project-nerds/ner-projectExpand on READMEUpdate cl cluster scripts for MLM finetuningUpdate finetuning scripts for cl clusterRevert to t5-base and finetuning fixesApply finetuning config made previously by kupper to MLM entity as wellAdd T5 MLM with entity masking, both classification and finetuning codeTest/Train split and finetuning config- forgot removing print for debug- READMEContext importance analysisClassification with Word2VecFinetuning for T5 MLMFinetune slurm script and loss plotNEC, fixed misunderstanding, removed unnecessary codeNLI with LLMs- GLiNER evaluation for all datasets except pile, with results (which are invalid for figer because it currently only returns one annotated entity for each sentence)Rename T5 to T5-NLI and save finetuned modelT5 MLM: Add next extra_id to end of label ids like it was done in training of T5 (see paper)Fix T5_MLM_labelAdd T5 MLM approach where the label is being maskedFiger: Add random samplingT5 NEC use highest probability labelFinished NEC evaluation (accuracy only)T5 NECNEC with GLiNER