<!-- 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-ted-id-en-lr_1e-3-ep_30-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: 1.5777
- Bleu: 18.3981
- Gen Len: 16.2277
- Meteor: 0.3652
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: 0.001
- 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 | Gen Len | Meteor |
---|---|---|---|---|---|---|
2.1196 | 1.0 | 1250 | 1.6942 | 13.7333 | 16.5568 | 0.3053 |
1.7331 | 2.0 | 2500 | 1.5393 | 15.6555 | 16.3938 | 0.3341 |
1.5842 | 3.0 | 3750 | 1.4821 | 16.2862 | 16.3699 | 0.3403 |
1.4518 | 4.0 | 5000 | 1.4562 | 17.1158 | 16.2073 | 0.3518 |
1.3649 | 5.0 | 6250 | 1.4416 | 17.3302 | 16.3344 | 0.355 |
1.2987 | 6.0 | 7500 | 1.4276 | 17.3334 | 16.3913 | 0.3547 |
1.2227 | 7.0 | 8750 | 1.4411 | 17.9415 | 16.2941 | 0.3612 |
1.1766 | 8.0 | 10000 | 1.4435 | 18.0776 | 16.2809 | 0.362 |
1.1119 | 9.0 | 11250 | 1.4510 | 18.0156 | 16.2834 | 0.3628 |
1.0672 | 10.0 | 12500 | 1.4566 | 18.276 | 16.1982 | 0.3646 |
1.0194 | 11.0 | 13750 | 1.4728 | 18.2417 | 16.2589 | 0.364 |
0.988 | 12.0 | 15000 | 1.4843 | 18.385 | 16.2671 | 0.3644 |
0.9445 | 13.0 | 16250 | 1.4896 | 18.1065 | 16.2321 | 0.3629 |
0.9156 | 14.0 | 17500 | 1.5102 | 18.249 | 16.2334 | 0.364 |
0.8888 | 15.0 | 18750 | 1.5273 | 18.1876 | 16.2454 | 0.3641 |
0.8624 | 16.0 | 20000 | 1.5354 | 18.2708 | 16.2425 | 0.3638 |
0.8364 | 17.0 | 21250 | 1.5449 | 18.17 | 16.268 | 0.3645 |
0.8133 | 18.0 | 22500 | 1.5578 | 18.2573 | 16.2422 | 0.3639 |
0.7994 | 19.0 | 23750 | 1.5678 | 18.3108 | 16.2158 | 0.3648 |
0.7823 | 20.0 | 25000 | 1.5777 | 18.3981 | 16.2277 | 0.3652 |
Framework versions
- Transformers 4.20.1
- Pytorch 1.11.0
- Datasets 2.1.0
- Tokenizers 0.12.1