BioNLP13CG_SciBERT_NER
This model is a fine-tuned version of allenai/scibert_scivocab_uncased on the None dataset. It achieves the following results on the evaluation set:
Loss: 0.1817
Seqeval classification report: precision recall f1-score support
Amino_acid 0.54 0.43 0.48 89 Anatomical_system 0.00 0.00 0.00 41 Cancer 0.84 0.84 0.84 3620 Cell 0.00 0.00 0.00 11 Cellular_component 0.00 0.00 0.00 7
Developing_anatomical_structure 0.00 0.00 0.00 37 Gene_or_gene_product 0.90 0.92 0.91 540 Immaterial_anatomical_entity 0.63 0.65 0.64 82 Multi-tissue_structure 0.63 0.71 0.67 144 Organ 0.00 0.00 0.00 56 Organism 0.86 0.17 0.28 36 Organism_subdivision 0.83 0.86 0.84 1086 Organism_substance 0.87 0.81 0.84 484 Pathological_formation 0.92 0.92 0.92 1430 Simple_chemical 0.58 0.72 0.64 304 Tissue 0.79 0.82 0.80 341
micro avg 0.84 0.82 0.83 8308
macro avg 0.52 0.49 0.49 8308
weighted avg 0.82 0.82 0.82 8308
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 | 0.2278 | precision recall f1-score support |
Amino_acid 0.48 0.15 0.22 89
Anatomical_system 0.00 0.00 0.00 41
Cancer 0.81 0.80 0.80 3620
Cell 0.00 0.00 0.00 11
Cellular_component 0.00 0.00 0.00 7
Developing_anatomical_structure 0.00 0.00 0.00 37 Gene_or_gene_product 0.80 0.90 0.84 540 Immaterial_anatomical_entity 0.48 0.59 0.52 82 Multi-tissue_structure 0.62 0.45 0.52 144 Organ 0.00 0.00 0.00 56 Organism 0.00 0.00 0.00 36 Organism_subdivision 0.75 0.84 0.79 1086 Organism_substance 0.83 0.77 0.80 484 Pathological_formation 0.90 0.86 0.88 1430 Simple_chemical 0.53 0.69 0.60 304 Tissue 0.74 0.73 0.73 341
micro avg 0.79 0.78 0.78 8308
macro avg 0.43 0.42 0.42 8308
weighted avg 0.77 0.78 0.77 8308
| | No log | 2.0 | 191 | 0.1850 | precision recall f1-score support
Amino_acid 0.52 0.40 0.46 89
Anatomical_system 0.00 0.00 0.00 41
Cancer 0.83 0.84 0.84 3620
Cell 0.00 0.00 0.00 11
Cellular_component 0.00 0.00 0.00 7
Developing_anatomical_structure 0.00 0.00 0.00 37 Gene_or_gene_product 0.89 0.92 0.90 540 Immaterial_anatomical_entity 0.56 0.65 0.60 82 Multi-tissue_structure 0.60 0.69 0.64 144 Organ 0.00 0.00 0.00 56 Organism 1.00 0.17 0.29 36 Organism_subdivision 0.80 0.87 0.83 1086 Organism_substance 0.87 0.79 0.83 484 Pathological_formation 0.91 0.93 0.92 1430 Simple_chemical 0.57 0.72 0.64 304 Tissue 0.77 0.79 0.78 341
micro avg 0.82 0.83 0.82 8308
macro avg 0.52 0.49 0.48 8308
weighted avg 0.81 0.83 0.82 8308
| | No log | 2.98 | 285 | 0.1817 | precision recall f1-score support
Amino_acid 0.54 0.43 0.48 89
Anatomical_system 0.00 0.00 0.00 41
Cancer 0.84 0.84 0.84 3620
Cell 0.00 0.00 0.00 11
Cellular_component 0.00 0.00 0.00 7
Developing_anatomical_structure 0.00 0.00 0.00 37 Gene_or_gene_product 0.90 0.92 0.91 540 Immaterial_anatomical_entity 0.63 0.65 0.64 82 Multi-tissue_structure 0.63 0.71 0.67 144 Organ 0.00 0.00 0.00 56 Organism 0.86 0.17 0.28 36 Organism_subdivision 0.83 0.86 0.84 1086 Organism_substance 0.87 0.81 0.84 484 Pathological_formation 0.92 0.92 0.92 1430 Simple_chemical 0.58 0.72 0.64 304 Tissue 0.79 0.82 0.80 341
micro avg 0.84 0.82 0.83 8308
macro avg 0.52 0.49 0.49 8308
weighted avg 0.82 0.82 0.82 8308
|
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
- Downloads last month
- 3