CeLLaTe 2-Class NER Models
Collection
This collection consolidates NER models trained using a 2-class label schema, predicting CellLine and Cell_Tissue (combined CellType and Tissue) • 9 items • Updated
CeLLaTe-ner-bioformer8l-baseline
This model is a fine-tuned version of bioformers/bioformer-8l on the OTAR3088/CeLLaTe_V3.2_contracted_ent_IOB dataset. It achieves the following results on the evaluation set:
- Loss: 0.0634
- Precision: 0.7493
- Recall: 0.7352
- F1: 0.7422
- Accuracy: 0.9810
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: 3407
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.2842 | 1.0 | 263 | 0.0718 | 0.5595 | 0.6185 | 0.5875 | 0.9732 |
| 0.0526 | 2.0 | 526 | 0.0693 | 0.6856 | 0.6047 | 0.6426 | 0.9761 |
| 0.0338 | 3.0 | 789 | 0.0705 | 0.6229 | 0.6221 | 0.6225 | 0.9741 |
| 0.0262 | 4.0 | 1052 | 0.0637 | 0.7493 | 0.7352 | 0.7422 | 0.9810 |
| 0.0197 | 5.0 | 1315 | 0.0800 | 0.7491 | 0.6852 | 0.7157 | 0.9791 |
| 0.0168 | 6.0 | 1578 | 0.0891 | 0.7522 | 0.6530 | 0.6991 | 0.9784 |
| 0.0129 | 7.0 | 1841 | 0.0797 | 0.7033 | 0.7357 | 0.7191 | 0.9791 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0
- Downloads last month
- 12