|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
- f1 |
|
model-index: |
|
- name: bert-nlp-project-ft-news-ds-google |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# bert-nlp-project-ft-news-ds-google |
|
|
|
This model is a fine-tuned version of [jestemleon/bert-nlp-project-news](https://huggingface.co/jestemleon/bert-nlp-project-news) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3238 |
|
- Accuracy: 0.9038 |
|
- F1: 0.9107 |
|
|
|
## 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 |
|
- 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 | Accuracy | F1 | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| |
|
| 0.3415 | 0.37 | 196 | 0.2452 | 0.9048 | 0.9136 | |
|
| 0.2747 | 0.75 | 392 | 0.2425 | 0.9076 | 0.9136 | |
|
| 0.2309 | 1.12 | 588 | 0.2658 | 0.9114 | 0.9179 | |
|
| 0.1964 | 1.49 | 784 | 0.2916 | 0.9010 | 0.9055 | |
|
| 0.1812 | 1.86 | 980 | 0.2950 | 0.9019 | 0.9091 | |
|
| 0.1569 | 2.24 | 1176 | 0.2926 | 0.9095 | 0.9164 | |
|
| 0.1173 | 2.61 | 1372 | 0.3188 | 0.9076 | 0.9145 | |
|
| 0.1145 | 2.98 | 1568 | 0.3238 | 0.9038 | 0.9107 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.25.1 |
|
- Pytorch 1.13.0+cu116 |
|
- Datasets 2.8.0 |
|
- Tokenizers 0.13.2 |
|
|
