base_model: medicalai/ClinicalBERT
tags:
- generated_from_trainer
model-index:
- name: BioNLP13CG_ClinicalBERT_NER
results: []
BioNLP13CG_ClinicalBERT_NER
This model is a fine-tuned version of medicalai/ClinicalBERT on the None dataset. It achieves the following results on the evaluation set:
Loss: 0.3339
Seqeval classification report: precision recall f1-score support
Amino_acid 0.81 0.59 0.68 297 Anatomical_system 0.70 0.78 0.74 297 Cancer 0.74 0.73 0.73 3490 Cell 0.72 0.87 0.79 1360 Cellular_component 0.00 0.00 0.00 99
Developing_anatomical_structure 0.00 0.00 0.00 11 Gene_or_gene_product 0.67 0.25 0.37 174 Immaterial_anatomical_entity 0.52 0.76 0.62 432 Multi-tissue_structure 0.83 0.59 0.69 317 Organ 0.00 0.00 0.00 49 Organism 0.71 0.48 0.57 464 Organism_subdivision 0.70 0.72 0.71 678 Organism_substance 0.00 0.00 0.00 128 Pathological_formation 0.62 0.05 0.09 108 Simple_chemical 0.00 0.00 0.00 56 Tissue 0.80 0.85 0.82 1566
micro avg 0.73 0.71 0.72 9526
macro avg 0.49 0.42 0.43 9526
weighted avg 0.71 0.71 0.70 9526
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 | Seqeval classification report |
|---|---|---|---|---|
| No log | 0.99 | 95 | 0.4681 | precision recall f1-score support |
Amino_acid 1.00 0.02 0.04 297
Anatomical_system 0.44 0.68 0.54 297
Cancer 0.68 0.63 0.65 3490
Cell 0.59 0.85 0.70 1360
Cellular_component 0.00 0.00 0.00 99
Developing_anatomical_structure 0.00 0.00 0.00 11 Gene_or_gene_product 0.00 0.00 0.00 174 Immaterial_anatomical_entity 0.40 0.60 0.48 432 Multi-tissue_structure 0.86 0.06 0.11 317 Organ 0.00 0.00 0.00 49 Organism 0.88 0.02 0.03 464 Organism_subdivision 0.62 0.54 0.58 678 Organism_substance 0.00 0.00 0.00 128 Pathological_formation 0.00 0.00 0.00 108 Simple_chemical 0.00 0.00 0.00 56 Tissue 0.70 0.84 0.76 1566
micro avg 0.63 0.58 0.60 9526
macro avg 0.39 0.27 0.24 9526
weighted avg 0.63 0.58 0.55 9526
| | No log | 2.0 | 191 | 0.3526 | precision recall f1-score support
Amino_acid 0.81 0.52 0.63 297
Anatomical_system 0.66 0.77 0.71 297
Cancer 0.74 0.73 0.73 3490
Cell 0.71 0.87 0.78 1360
Cellular_component 0.00 0.00 0.00 99
Developing_anatomical_structure 0.00 0.00 0.00 11 Gene_or_gene_product 0.76 0.20 0.32 174 Immaterial_anatomical_entity 0.46 0.76 0.57 432 Multi-tissue_structure 0.83 0.57 0.68 317 Organ 0.00 0.00 0.00 49 Organism 0.68 0.44 0.54 464 Organism_subdivision 0.71 0.67 0.69 678 Organism_substance 0.00 0.00 0.00 128 Pathological_formation 1.00 0.01 0.02 108 Simple_chemical 0.00 0.00 0.00 56 Tissue 0.78 0.85 0.81 1566
micro avg 0.72 0.70 0.71 9526
macro avg 0.51 0.40 0.41 9526
weighted avg 0.70 0.70 0.68 9526
| | No log | 2.98 | 285 | 0.3339 | precision recall f1-score support
Amino_acid 0.81 0.59 0.68 297
Anatomical_system 0.70 0.78 0.74 297
Cancer 0.74 0.73 0.73 3490
Cell 0.72 0.87 0.79 1360
Cellular_component 0.00 0.00 0.00 99
Developing_anatomical_structure 0.00 0.00 0.00 11 Gene_or_gene_product 0.67 0.25 0.37 174 Immaterial_anatomical_entity 0.52 0.76 0.62 432 Multi-tissue_structure 0.83 0.59 0.69 317 Organ 0.00 0.00 0.00 49 Organism 0.71 0.48 0.57 464 Organism_subdivision 0.70 0.72 0.71 678 Organism_substance 0.00 0.00 0.00 128 Pathological_formation 0.62 0.05 0.09 108 Simple_chemical 0.00 0.00 0.00 56 Tissue 0.80 0.85 0.82 1566
micro avg 0.73 0.71 0.72 9526
macro avg 0.49 0.42 0.43 9526
weighted avg 0.71 0.71 0.70 9526
|
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0