NLLB-alt-cv-bleu-40
This model is a fine-tuned version of facebook/nllb-200-distilled-600M on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.7734
- Bleu: 30.3568
- Gen Len: 50.699
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 40
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
---|---|---|---|---|---|
2.1031 | 1.0 | 1380 | 1.6725 | 6.1774 | 56.292 |
1.4925 | 2.0 | 2760 | 1.3799 | 8.9414 | 54.966 |
1.1976 | 3.0 | 4140 | 1.2417 | 10.7993 | 55.32 |
0.9973 | 4.0 | 5520 | 1.1744 | 13.7633 | 51.225 |
0.8305 | 5.0 | 6900 | 1.1461 | 14.8273 | 51.723 |
0.6918 | 6.0 | 8280 | 1.1261 | 16.02 | 50.83 |
0.5823 | 7.0 | 9660 | 1.1556 | 17.8129 | 49.93 |
0.4934 | 8.0 | 11040 | 1.1567 | 19.3213 | 50.647 |
0.4049 | 9.0 | 12420 | 1.1794 | 21.712 | 50.722 |
0.3342 | 10.0 | 13800 | 1.2320 | 22.2956 | 50.552 |
0.2747 | 11.0 | 15180 | 1.2749 | 24.3631 | 50.443 |
0.2281 | 12.0 | 16560 | 1.2996 | 25.4046 | 52.437 |
0.1849 | 13.0 | 17940 | 1.3378 | 26.5399 | 50.361 |
0.153 | 14.0 | 19320 | 1.3709 | 27.0563 | 51.077 |
0.1256 | 15.0 | 20700 | 1.4128 | 27.8781 | 51.129 |
0.1103 | 16.0 | 22080 | 1.4354 | 28.6894 | 51.974 |
0.0893 | 17.0 | 23460 | 1.4859 | 28.0852 | 52.005 |
0.0778 | 18.0 | 24840 | 1.4973 | 28.9053 | 50.803 |
0.0683 | 19.0 | 26220 | 1.5294 | 29.2219 | 50.845 |
0.0592 | 20.0 | 27600 | 1.5576 | 29.1227 | 51.051 |
0.0505 | 21.0 | 28980 | 1.5885 | 29.4121 | 50.376 |
0.0441 | 22.0 | 30360 | 1.6028 | 29.5531 | 51.946 |
0.0397 | 23.0 | 31740 | 1.6254 | 29.3607 | 50.811 |
0.0361 | 24.0 | 33120 | 1.6374 | 29.5197 | 51.166 |
0.0323 | 25.0 | 34500 | 1.6423 | 29.7589 | 51.335 |
0.0288 | 26.0 | 35880 | 1.6630 | 29.6029 | 51.036 |
0.0257 | 27.0 | 37260 | 1.6800 | 29.4437 | 50.623 |
0.0235 | 28.0 | 38640 | 1.6887 | 29.9344 | 50.797 |
0.0201 | 29.0 | 40020 | 1.7096 | 30.1522 | 50.694 |
0.018 | 30.0 | 41400 | 1.7223 | 30.1291 | 50.425 |
0.0163 | 31.0 | 42780 | 1.7282 | 29.8131 | 51.114 |
0.0148 | 32.0 | 44160 | 1.7299 | 29.9721 | 50.851 |
0.0133 | 33.0 | 45540 | 1.7463 | 30.0369 | 50.477 |
0.0122 | 34.0 | 46920 | 1.7514 | 30.0663 | 51.133 |
0.0112 | 35.0 | 48300 | 1.7508 | 30.0451 | 50.736 |
0.0099 | 36.0 | 49680 | 1.7631 | 30.0576 | 50.62 |
0.0087 | 37.0 | 51060 | 1.7683 | 30.1648 | 50.874 |
0.0083 | 38.0 | 52440 | 1.7750 | 30.2558 | 50.667 |
0.0076 | 39.0 | 53820 | 1.7757 | 30.3551 | 50.886 |
0.0076 | 40.0 | 55200 | 1.7734 | 30.3568 | 50.699 |
Framework versions
- Transformers 4.21.0
- Pytorch 1.10.0+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
- Downloads last month
- 15
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.