<!-- 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. -->
nmt-mpst-id-en-lr_0.0001-ep_20-seq_128_bs-16
This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.8531
- Bleu: 0.1306
- Meteor: 0.2859
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: 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: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Meteor |
|---|---|---|---|---|---|
| No log | 1.0 | 404 | 2.7171 | 0.0374 | 0.14 |
| 3.1222 | 2.0 | 808 | 2.4821 | 0.0519 | 0.1723 |
| 2.7305 | 3.0 | 1212 | 2.3370 | 0.0663 | 0.1983 |
| 2.4848 | 4.0 | 1616 | 2.2469 | 0.0771 | 0.2158 |
| 2.3394 | 5.0 | 2020 | 2.1567 | 0.0857 | 0.227 |
| 2.3394 | 6.0 | 2424 | 2.1038 | 0.0919 | 0.2369 |
| 2.2007 | 7.0 | 2828 | 2.0403 | 0.0973 | 0.2449 |
| 2.1027 | 8.0 | 3232 | 2.0105 | 0.1066 | 0.2554 |
| 2.0299 | 9.0 | 3636 | 1.9725 | 0.1105 | 0.2606 |
| 1.9568 | 10.0 | 4040 | 1.9515 | 0.1147 | 0.2655 |
| 1.9568 | 11.0 | 4444 | 1.9274 | 0.118 | 0.2699 |
| 1.8986 | 12.0 | 4848 | 1.9142 | 0.1215 | 0.2739 |
| 1.8512 | 13.0 | 5252 | 1.8936 | 0.1243 | 0.2777 |
| 1.8258 | 14.0 | 5656 | 1.8841 | 0.1254 | 0.279 |
| 1.7854 | 15.0 | 6060 | 1.8792 | 0.1278 | 0.2827 |
| 1.7854 | 16.0 | 6464 | 1.8662 | 0.1274 | 0.2818 |
| 1.7598 | 17.0 | 6868 | 1.8604 | 0.1293 | 0.2834 |
| 1.7436 | 18.0 | 7272 | 1.8598 | 0.13 | 0.2849 |
| 1.7299 | 19.0 | 7676 | 1.8545 | 0.1308 | 0.2857 |
| 1.7168 | 20.0 | 8080 | 1.8531 | 0.1306 | 0.2859 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Datasets 2.7.0
- Tokenizers 0.13.2
