metadata
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: cantonese-chinese-translation
results: []
cantonese-chinese-translation
This model is a fine-tuned version of fnlp/bart-base-chinese on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2370
- Bleu: 62.9641
- Chrf: 60.8644
- Gen Len: 12.8539
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: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Chrf | Gen Len |
|---|---|---|---|---|---|---|
| 0.3589 | 0.48 | 1000 | 0.2446 | 62.6737 | 60.5633 | 12.874 |
| 0.3038 | 0.96 | 2000 | 0.2347 | 62.8493 | 60.8189 | 12.8479 |
| 0.2301 | 1.44 | 3000 | 0.2370 | 62.9641 | 60.8644 | 12.8539 |
| 0.2267 | 1.92 | 4000 | 0.2323 | 62.9373 | 60.7878 | 12.9001 |
| 0.1725 | 2.39 | 5000 | 0.2431 | 62.7294 | 60.5396 | 12.8728 |
| 0.1665 | 2.87 | 6000 | 0.2399 | 62.6086 | 60.4485 | 12.8479 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.13.3