marianmt_sanlish_to_bangla_1934_lr2
This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-mul on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.5286
- Bleu: 53.7285
- Chrf: 73.2557
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 25
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Chrf |
|---|---|---|---|---|---|
| 2.8256 | 1.0 | 196 | 2.0909 | 15.1069 | 26.5352 |
| 2.0816 | 2.0 | 392 | 1.8137 | 11.4787 | 23.8366 |
| 1.7841 | 3.0 | 588 | 1.6477 | 20.4124 | 48.0147 |
| 1.5486 | 4.0 | 784 | 1.5554 | 25.4066 | 46.6072 |
| 1.3411 | 5.0 | 980 | 1.5086 | 25.4066 | 43.2104 |
| 1.1955 | 6.0 | 1176 | 1.5009 | 35.3553 | 47.3758 |
| 1.0821 | 7.0 | 1372 | 1.4746 | 35.3553 | 50.4288 |
| 0.9838 | 8.0 | 1568 | 1.4638 | 35.3553 | 50.4288 |
| 0.8870 | 9.0 | 1764 | 1.4497 | 17.9652 | 57.7094 |
| 0.7993 | 10.0 | 1960 | 1.4479 | 17.9652 | 57.7094 |
| 0.7311 | 11.0 | 2156 | 1.4527 | 50.8133 | 68.3656 |
| 0.6652 | 12.0 | 2352 | 1.4618 | 50.8133 | 68.3656 |
| 0.6177 | 13.0 | 2548 | 1.4612 | 50.8133 | 75.7976 |
| 0.5655 | 14.0 | 2744 | 1.4740 | 50.8133 | 75.7976 |
| 0.5311 | 15.0 | 2940 | 1.4883 | 50.8133 | 68.3656 |
| 0.4830 | 16.0 | 3136 | 1.4928 | 50.8133 | 68.3656 |
| 0.4528 | 17.0 | 3332 | 1.5003 | 50.8133 | 68.3656 |
| 0.4296 | 18.0 | 3528 | 1.4984 | 50.8133 | 67.0024 |
| 0.4063 | 19.0 | 3724 | 1.5172 | 50.8133 | 67.0024 |
| 0.3840 | 20.0 | 3920 | 1.5165 | 50.8133 | 67.0024 |
| 0.3761 | 21.0 | 4116 | 1.5189 | 50.8133 | 67.0024 |
| 0.3571 | 22.0 | 4312 | 1.5259 | 50.8133 | 67.0024 |
| 0.3429 | 23.0 | 4508 | 1.5270 | 53.7285 | 73.2557 |
| 0.3399 | 24.0 | 4704 | 1.5288 | 53.7285 | 73.2557 |
| 0.3366 | 25.0 | 4900 | 1.5286 | 53.7285 | 73.2557 |
Framework versions
- Transformers 5.15.0
- Pytorch 2.11.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2