multiCorp_2e-05_0404
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: 0.0717
- Precision: 0.5887
- Recall: 0.5283
- F1: 0.5569
- Accuracy: 0.9834
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: 2000
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
1.4744 | 0.08 | 25 | 0.2177 | 0.0 | 0.0 | 0.0 | 0.9735 |
0.1749 | 0.15 | 50 | 0.2022 | 0.0 | 0.0 | 0.0 | 0.9735 |
0.1853 | 0.23 | 75 | 0.2006 | 0.0 | 0.0 | 0.0 | 0.9735 |
0.188 | 0.31 | 100 | 0.1984 | 0.0 | 0.0 | 0.0 | 0.9735 |
0.1662 | 0.39 | 125 | 0.1869 | 0.0 | 0.0 | 0.0 | 0.9735 |
0.1824 | 0.46 | 150 | 0.1656 | 0.0 | 0.0 | 0.0 | 0.9735 |
0.1672 | 0.54 | 175 | 0.1443 | 0.8214 | 0.0107 | 0.0211 | 0.9737 |
0.1269 | 0.62 | 200 | 0.1296 | 0.2189 | 0.1110 | 0.1473 | 0.9740 |
0.116 | 0.7 | 225 | 0.1221 | 0.3206 | 0.1982 | 0.2450 | 0.9753 |
0.111 | 0.77 | 250 | 0.1208 | 0.4109 | 0.1968 | 0.2662 | 0.9762 |
0.1382 | 0.85 | 275 | 0.1149 | 0.4680 | 0.1495 | 0.2266 | 0.9763 |
0.1136 | 0.93 | 300 | 0.1051 | 0.3749 | 0.2303 | 0.2853 | 0.9767 |
0.1043 | 1.01 | 325 | 0.1066 | 0.3451 | 0.3315 | 0.3381 | 0.9762 |
0.1062 | 1.08 | 350 | 0.1012 | 0.4072 | 0.3319 | 0.3657 | 0.9769 |
0.0834 | 1.16 | 375 | 0.0972 | 0.4079 | 0.3301 | 0.3649 | 0.9781 |
0.0923 | 1.24 | 400 | 0.0973 | 0.4598 | 0.3500 | 0.3975 | 0.9781 |
0.0932 | 1.32 | 425 | 0.0932 | 0.4649 | 0.3384 | 0.3917 | 0.9789 |
0.1044 | 1.39 | 450 | 0.0934 | 0.5039 | 0.3319 | 0.4002 | 0.9792 |
0.0962 | 1.47 | 475 | 0.0926 | 0.4636 | 0.3045 | 0.3676 | 0.9788 |
0.079 | 1.55 | 500 | 0.0883 | 0.4772 | 0.3890 | 0.4286 | 0.9799 |
0.0792 | 1.63 | 525 | 0.0856 | 0.4520 | 0.3890 | 0.4182 | 0.9799 |
0.0823 | 1.7 | 550 | 0.0847 | 0.4618 | 0.4517 | 0.4567 | 0.9799 |
0.079 | 1.78 | 575 | 0.0830 | 0.5208 | 0.3890 | 0.4454 | 0.9805 |
0.0832 | 1.86 | 600 | 0.0830 | 0.5201 | 0.3538 | 0.4211 | 0.9803 |
0.0688 | 1.93 | 625 | 0.0824 | 0.4816 | 0.4550 | 0.4679 | 0.9806 |
0.0752 | 2.01 | 650 | 0.0786 | 0.4956 | 0.4401 | 0.4662 | 0.9810 |
0.0699 | 2.09 | 675 | 0.0795 | 0.5304 | 0.4698 | 0.4983 | 0.9817 |
0.0705 | 2.17 | 700 | 0.0777 | 0.4963 | 0.4954 | 0.4958 | 0.9813 |
0.0591 | 2.24 | 725 | 0.0807 | 0.5545 | 0.4438 | 0.4930 | 0.9818 |
0.0641 | 2.32 | 750 | 0.0793 | 0.5270 | 0.4257 | 0.4710 | 0.9814 |
0.0761 | 2.4 | 775 | 0.0745 | 0.5150 | 0.4796 | 0.4966 | 0.9818 |
0.068 | 2.48 | 800 | 0.0765 | 0.5741 | 0.4262 | 0.4892 | 0.9819 |
0.0596 | 2.55 | 825 | 0.0757 | 0.5346 | 0.4341 | 0.4791 | 0.9817 |
0.0648 | 2.63 | 850 | 0.0724 | 0.5526 | 0.5023 | 0.5263 | 0.9827 |
0.0619 | 2.71 | 875 | 0.0739 | 0.5471 | 0.5288 | 0.5378 | 0.9824 |
0.06 | 2.79 | 900 | 0.0738 | 0.5627 | 0.5227 | 0.5420 | 0.9829 |
0.058 | 2.86 | 925 | 0.0740 | 0.5456 | 0.5107 | 0.5276 | 0.9825 |
0.0624 | 2.94 | 950 | 0.0712 | 0.5665 | 0.5237 | 0.5443 | 0.9832 |
0.0602 | 3.02 | 975 | 0.0700 | 0.5368 | 0.5181 | 0.5273 | 0.9828 |
0.049 | 3.1 | 1000 | 0.0720 | 0.5710 | 0.5339 | 0.5518 | 0.9832 |
0.0562 | 3.17 | 1025 | 0.0715 | 0.5847 | 0.5176 | 0.5491 | 0.9831 |
0.0559 | 3.25 | 1050 | 0.0711 | 0.5921 | 0.5460 | 0.5681 | 0.9834 |
0.054 | 3.33 | 1075 | 0.0707 | 0.6062 | 0.5395 | 0.5709 | 0.9837 |
0.0522 | 3.41 | 1100 | 0.0716 | 0.5530 | 0.5209 | 0.5365 | 0.9828 |
0.0456 | 3.48 | 1125 | 0.0717 | 0.5887 | 0.5283 | 0.5569 | 0.9834 |
Framework versions
- Transformers 4.27.4
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.2
- Downloads last month
- 10
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.