CRAFT_SciBERT_NER

• Loss: 0.1143
• Seqeval classification report: precision recall f1-score support

   CHEBI       0.74      0.70      0.72       457
      CL       0.82      0.75      0.78      1099
     GGP       0.92      0.93      0.93      2232
      GO       0.78      0.84      0.81      2508
      SO       0.83      0.81      0.82      1365
   Taxon       0.99      0.99      0.99     87655

  micro avg 0.98 0.98 0.98 95316 macro avg 0.85 0.84 0.84 95316 weighted avg 0.98 0.98 0.98 95316

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

   CHEBI       0.66      0.69      0.67       457
      CL       0.83      0.69      0.75      1099
     GGP       0.89      0.93      0.91      2232
      GO       0.76      0.85      0.80      2508
      SO       0.79      0.73      0.76      1365
   Taxon       0.99      0.99      0.99     87655

   CHEBI       0.73      0.69      0.71       457
      CL       0.85      0.72      0.78      1099
     GGP       0.91      0.93      0.92      2232
      GO       0.74      0.87      0.80      2508
      SO       0.82      0.80      0.81      1365
   Taxon       0.99      0.99      0.99     87655

   CHEBI       0.74      0.70      0.72       457
      CL       0.82      0.75      0.78      1099
     GGP       0.92      0.93      0.93      2232
      GO       0.78      0.84      0.81      2508
      SO       0.83      0.81      0.82      1365
   Taxon       0.99      0.99      0.99     87655

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0