<!-- 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. -->
fugumt-ja-en-finetuned-ja-to-en
This model is a fine-tuned version of staka/fugumt-ja-en on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1447
- Bleu: 86.2696
- Gen Len: 6.5991
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: 25
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 4.5109 | 1.0 | 544 | 1.3336 | 22.852 | 8.2346 |
| 1.4507 | 2.0 | 1088 | 1.0665 | 32.3126 | 7.7839 |
| 1.2319 | 3.0 | 1632 | 0.8913 | 39.464 | 7.324 |
| 1.0674 | 4.0 | 2176 | 0.7552 | 44.1825 | 7.2759 |
| 0.9451 | 5.0 | 2720 | 0.6529 | 50.7008 | 6.8755 |
| 0.8415 | 6.0 | 3264 | 0.5657 | 54.1043 | 7.0116 |
| 0.758 | 7.0 | 3808 | 0.4963 | 54.6872 | 7.3108 |
| 0.6901 | 8.0 | 4352 | 0.4373 | 57.4437 | 7.2211 |
| 0.621 | 9.0 | 4896 | 0.3878 | 56.9356 | 7.2811 |
| 0.5729 | 10.0 | 5440 | 0.3464 | 65.85 | 6.8777 |
| 0.5311 | 11.0 | 5984 | 0.3112 | 68.8388 | 6.7775 |
| 0.4463 | 12.0 | 6528 | 0.2812 | 65.2178 | 7.2367 |
| 0.4182 | 13.0 | 7072 | 0.2569 | 76.2492 | 6.5945 |
| 0.3893 | 14.0 | 7616 | 0.2346 | 75.3988 | 6.8026 |
| 0.3713 | 15.0 | 8160 | 0.2167 | 78.6206 | 6.7092 |
| 0.3469 | 16.0 | 8704 | 0.2010 | 79.4506 | 6.7256 |
| 0.3304 | 17.0 | 9248 | 0.1881 | 81.9813 | 6.5934 |
| 0.3162 | 18.0 | 9792 | 0.1784 | 82.5415 | 6.5888 |
| 0.3022 | 19.0 | 10336 | 0.1685 | 83.2969 | 6.6102 |
| 0.29 | 20.0 | 10880 | 0.1614 | 83.7095 | 6.6235 |
| 0.2783 | 21.0 | 11424 | 0.1560 | 83.683 | 6.6559 |
| 0.27 | 22.0 | 11968 | 0.1508 | 84.719 | 6.6072 |
| 0.2594 | 23.0 | 12512 | 0.1475 | 87.3934 | 6.5284 |
| 0.2469 | 24.0 | 13056 | 0.1455 | 85.2872 | 6.6427 |
| 0.2492 | 25.0 | 13600 | 0.1447 | 86.2696 | 6.5991 |
Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
