metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: opus-mt-en-ru-finetuned
results: []
opus-mt-en-ru-finetuned
This model is a fine-tuned version of kazandaev/opus-mt-en-ru-finetuned on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.8061
- Bleu: 39.3089
- Gen Len: 29.6895
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: 49
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 0.8593 | 1.0 | 35147 | 0.8723 | 37.1365 | 29.6618 |
| 0.8611 | 2.0 | 70294 | 0.8582 | 37.6328 | 29.7661 |
| 0.8825 | 3.0 | 105441 | 0.8529 | 37.2852 | 29.9074 |
| 0.8644 | 4.0 | 140588 | 0.8432 | 38.088 | 29.632 |
| 0.8555 | 5.0 | 175735 | 0.8341 | 37.9338 | 29.8584 |
| 0.8492 | 6.0 | 210882 | 0.8281 | 38.4253 | 29.7771 |
| 0.845 | 7.0 | 246029 | 0.8190 | 38.8127 | 29.818 |
| 0.8338 | 8.0 | 281176 | 0.8080 | 39.2423 | 29.7133 |
| 0.8327 | 9.0 | 316323 | 0.8080 | 39.082 | 29.7484 |
| 0.8396 | 10.0 | 351470 | 0.8061 | 39.3089 | 29.6895 |
Framework versions
- Transformers 4.16.2
- Pytorch 1.10.0+cu111
- Datasets 1.18.3
- Tokenizers 0.11.0