metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: MarianMix_en-10
results: []
MarianMix_en-10
This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-zh on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.0003
- Bleu: 15.2309
- Gen Len: 36.7472
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: 0.0001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 99
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 1.7988 | 0.44 | 500 | 1.9422 | 0.0 | 8.1057 |
| 1.0094 | 0.89 | 1000 | 1.6961 | 0.0044 | 11.6113 |
| 0.849 | 1.33 | 1500 | 1.4661 | 5.0612 | 42.1319 |
| 0.7512 | 1.78 | 2000 | 1.2904 | 7.2949 | 41.1219 |
| 0.6542 | 2.22 | 2500 | 1.1847 | 10.0266 | 41.9235 |
| 0.6081 | 2.67 | 3000 | 1.1217 | 9.5742 | 33.0483 |
| 0.5694 | 3.11 | 3500 | 1.0763 | 13.6519 | 40.5599 |
| 0.5304 | 3.56 | 4000 | 1.0447 | 13.6182 | 37.2709 |
| 0.5171 | 4.0 | 4500 | 1.0193 | 14.4302 | 38.7533 |
| 0.4841 | 4.44 | 5000 | 1.0093 | 14.898 | 36.4763 |
| 0.4779 | 4.89 | 5500 | 1.0003 | 15.2309 | 36.7472 |
Framework versions
- Transformers 4.12.5
- Pytorch 1.9.1
- Datasets 1.17.0
- Tokenizers 0.10.3