Edit model card

multiCorp_cut_5e-05_250

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: 0.0722
  • Precision: 0.7331
  • Recall: 0.6101
  • F1: 0.6660
  • Accuracy: 0.9851

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: 5e-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: 1500

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.2942 0.66 50 0.0969 0.4120 0.1031 0.1649 0.9742
0.0781 1.32 100 0.0720 0.4497 0.2933 0.3551 0.9779
0.0673 1.97 150 0.0632 0.5067 0.4227 0.4609 0.9794
0.046 2.63 200 0.0617 0.5822 0.4780 0.5250 0.9814
0.0426 3.29 250 0.0607 0.6815 0.5033 0.5790 0.9833
0.0365 3.95 300 0.0596 0.6222 0.5558 0.5871 0.9823
0.026 4.61 350 0.0632 0.6341 0.6026 0.6180 0.9829
0.0224 5.26 400 0.0700 0.6951 0.5576 0.6188 0.9838
0.0197 5.92 450 0.0621 0.6834 0.5989 0.6384 0.9840
0.0147 6.58 500 0.0676 0.7181 0.6064 0.6575 0.9845
0.0119 7.24 550 0.0689 0.6965 0.6560 0.6757 0.9846
0.0106 7.89 600 0.0722 0.7331 0.6101 0.6660 0.9851

Framework versions

  • Transformers 4.27.4
  • Pytorch 1.13.1+cu116
  • Datasets 2.11.0
  • Tokenizers 0.13.2
Downloads last month
9