metadata
license: apache-2.0
base_model: google-bert/bert-base-cased
tags:
- generated_from_trainer
metrics:
- f1
- recall
model-index:
- name: bert-base-cased
results: []
bert-base-cased
This model is a fine-tuned version of google-bert/bert-base-cased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3765
- F1 Macro: 0.8988
- F1: 0.9360
- F1 Neg: 0.8617
- Acc: 0.9125
- Prec: 0.9209
- Recall: 0.9517
- Mcc: 0.7989
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
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Macro | F1 | F1 Neg | Acc | Prec | Recall | Mcc |
|---|---|---|---|---|---|---|---|---|---|---|
| 0.5022 | 1.0 | 824 | 0.4717 | 0.8752 | 0.9102 | 0.8403 | 0.885 | 0.9102 | 0.9102 | 0.7504 |
| 0.3452 | 2.0 | 1648 | 0.4330 | 0.8882 | 0.9245 | 0.8519 | 0.9 | 0.8942 | 0.9570 | 0.7808 |
| 0.2232 | 3.0 | 2472 | 0.5604 | 0.8652 | 0.9060 | 0.8244 | 0.8775 | 0.8906 | 0.9219 | 0.7314 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2