metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- ncbi_disease
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: bert-finetuned-ncbi
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: ncbi_disease
type: ncbi_disease
config: ncbi_disease
split: train
args: ncbi_disease
metrics:
- name: Precision
type: precision
value: 0.7807118254879449
- name: Recall
type: recall
value: 0.8640406607369758
- name: F1
type: f1
value: 0.8202653799758745
- name: Accuracy
type: accuracy
value: 0.9831009585459978
bert-finetuned-ncbi
This model is a fine-tuned version of bert-base-cased on the ncbi_disease dataset. It achieves the following results on the evaluation set:
- Loss: 0.0679
- Precision: 0.7807
- Recall: 0.8640
- F1: 0.8203
- Accuracy: 0.9831
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: 8
- eval_batch_size: 8
- seed: 42
- 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 |
|---|---|---|---|---|---|---|---|
| 0.1146 | 1.0 | 680 | 0.0686 | 0.7450 | 0.8056 | 0.7741 | 0.9805 |
| 0.0458 | 2.0 | 1360 | 0.0612 | 0.7646 | 0.8628 | 0.8107 | 0.9815 |
| 0.0161 | 3.0 | 2040 | 0.0679 | 0.7807 | 0.8640 | 0.8203 | 0.9831 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.1+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2