metadata
license: mit
base_model: microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: PubMedBERT_CRAFT_NER
results: []
PubMedBERT_CRAFT_NER
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: 0.1031
- Precision: 0.8429
- Recall: 0.8679
- F1: 0.8552
- Accuracy: 0.9734
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
- 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
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 347 | 0.1280 | 0.7851 | 0.8360 | 0.8097 | 0.9647 |
| 0.1944 | 2.0 | 695 | 0.1092 | 0.8187 | 0.8615 | 0.8395 | 0.9707 |
| 0.046 | 3.0 | 1041 | 0.1031 | 0.8429 | 0.8679 | 0.8552 | 0.9734 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.0
- Tokenizers 0.15.0