<!-- 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-32
This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.9722
- Bleu: 0.1118
- Meteor: 0.2641
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: 32
- eval_batch_size: 32
- 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 | 202 | 2.8076 | 0.0316 | 0.1256 |
| No log | 2.0 | 404 | 2.6213 | 0.0427 | 0.1545 |
| 3.0404 | 3.0 | 606 | 2.4851 | 0.0532 | 0.1754 |
| 3.0404 | 4.0 | 808 | 2.3880 | 0.0605 | 0.1894 |
| 2.5973 | 5.0 | 1010 | 2.3137 | 0.0685 | 0.2014 |
| 2.5973 | 6.0 | 1212 | 2.2489 | 0.0729 | 0.2084 |
| 2.5973 | 7.0 | 1414 | 2.1949 | 0.0798 | 0.2199 |
| 2.3553 | 8.0 | 1616 | 2.1503 | 0.0854 | 0.227 |
| 2.3553 | 9.0 | 1818 | 2.1173 | 0.0915 | 0.2357 |
| 2.2044 | 10.0 | 2020 | 2.0854 | 0.0938 | 0.2397 |
| 2.2044 | 11.0 | 2222 | 2.0586 | 0.0974 | 0.2442 |
| 2.2044 | 12.0 | 2424 | 2.0418 | 0.1007 | 0.2491 |
| 2.0911 | 13.0 | 2626 | 2.0239 | 0.1033 | 0.2528 |
| 2.0911 | 14.0 | 2828 | 2.0071 | 0.105 | 0.255 |
| 2.0255 | 15.0 | 3030 | 1.9955 | 0.1068 | 0.2576 |
| 2.0255 | 16.0 | 3232 | 1.9913 | 0.1089 | 0.2609 |
| 2.0255 | 17.0 | 3434 | 1.9774 | 0.1099 | 0.2605 |
| 1.9777 | 18.0 | 3636 | 1.9789 | 0.1114 | 0.2638 |
| 1.9777 | 19.0 | 3838 | 1.9734 | 0.1116 | 0.2638 |
| 1.9505 | 20.0 | 4040 | 1.9722 | 0.1118 | 0.2641 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Datasets 2.7.0
- Tokenizers 0.13.2
