Edit model card

tmvar_2e-05_ES2

This model is a fine-tuned version of microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0184
  • Precision: 0.8368
  • Recall: 0.8595
  • F1: 0.848
  • Accuracy: 0.9962

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 1000

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.5018 1.47 25 0.1002 0.0 0.0 0.0 0.9843
0.0852 2.94 50 0.0509 0.9286 0.0703 0.1307 0.9852
0.0373 4.41 75 0.0283 0.5485 0.6108 0.5780 0.9918
0.0256 5.88 100 0.0204 0.6429 0.7297 0.6835 0.9938
0.0123 7.35 125 0.0188 0.8063 0.8324 0.8191 0.9956
0.008 8.82 150 0.0171 0.7979 0.8324 0.8148 0.9958
0.0047 10.29 175 0.0158 0.8010 0.8919 0.8440 0.9962
0.0037 11.76 200 0.0171 0.8511 0.8649 0.8579 0.9964
0.0025 13.24 225 0.0184 0.8368 0.8595 0.848 0.9962

Framework versions

  • Transformers 4.27.4
  • Pytorch 2.0.0+cu118
  • Datasets 2.11.0
  • Tokenizers 0.13.2
Downloads last month
8