|
--- |
|
license: apache-2.0 |
|
base_model: Helsinki-NLP/opus-mt-en-zh |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- bleu |
|
model-index: |
|
- name: opus-mt-cantonese-v1 |
|
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. --> |
|
|
|
# opus-mt-cantonese-v1 |
|
|
|
This model is a fine-tuned version of [Helsinki-NLP/opus-mt-en-zh](https://huggingface.co/Helsinki-NLP/opus-mt-en-zh) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 3.7154 |
|
- Bleu: 1.4102 |
|
- Gen Len: 12.6992 |
|
|
|
## 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-06 |
|
- 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: 10 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:-------:| |
|
| No log | 1.0 | 62 | 3.7154 | 1.3724 | 12.7073 | |
|
| No log | 2.0 | 124 | 3.7185 | 1.3718 | 12.7398 | |
|
| No log | 3.0 | 186 | 3.7168 | 1.4342 | 12.6504 | |
|
| No log | 4.0 | 248 | 3.7168 | 1.4109 | 12.7805 | |
|
| No log | 5.0 | 310 | 3.7161 | 1.4265 | 12.7073 | |
|
| No log | 6.0 | 372 | 3.7150 | 1.4013 | 12.7114 | |
|
| No log | 7.0 | 434 | 3.7151 | 1.4102 | 12.7073 | |
|
| No log | 8.0 | 496 | 3.7152 | 1.3981 | 12.7439 | |
|
| 2.0086 | 9.0 | 558 | 3.7152 | 1.4327 | 12.6951 | |
|
| 2.0086 | 10.0 | 620 | 3.7154 | 1.4102 | 12.6992 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.39.2 |
|
- Pytorch 2.2.1+cu121 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |
|
|
