metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- ccmatrix
metrics:
- bleu
model-index:
- name: t5-small-finetuned-en-to-it
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: ccmatrix
type: ccmatrix
config: en-it
split: train[3000:15000]
args: en-it
metrics:
- name: Bleu
type: bleu
value: 7.7504
t5-small-finetuned-en-to-it
This model is a fine-tuned version of t5-small on the ccmatrix dataset. It achieves the following results on the evaluation set:
- Loss: 2.2736
- Bleu: 7.7504
- Gen Len: 60.0173
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: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 3.4319 | 1.0 | 750 | 2.8398 | 2.2549 | 92.1093 |
| 3.0931 | 2.0 | 1500 | 2.7018 | 3.061 | 87.6667 |
| 3.0119 | 3.0 | 2250 | 2.6072 | 3.8085 | 79.9573 |
| 2.8924 | 4.0 | 3000 | 2.5345 | 4.4824 | 73.2513 |
| 2.8415 | 5.0 | 3750 | 2.4761 | 5.0326 | 70.2667 |
| 2.7656 | 6.0 | 4500 | 2.4320 | 5.4965 | 67.55 |
| 2.7305 | 7.0 | 5250 | 2.3915 | 6.1202 | 65.9733 |
| 2.6847 | 8.0 | 6000 | 2.3605 | 6.4886 | 64.696 |
| 2.656 | 9.0 | 6750 | 2.3352 | 6.7504 | 62.7593 |
| 2.6252 | 10.0 | 7500 | 2.3161 | 7.1305 | 61.516 |
| 2.6101 | 11.0 | 8250 | 2.3001 | 7.2954 | 61.1827 |
| 2.5974 | 12.0 | 9000 | 2.2882 | 7.516 | 60.974 |
| 2.5815 | 13.0 | 9750 | 2.2798 | 7.6634 | 60.4747 |
| 2.5665 | 14.0 | 10500 | 2.2750 | 7.6801 | 60.3567 |
| 2.5688 | 15.0 | 11250 | 2.2736 | 7.7504 | 60.0173 |
Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.5.1
- Tokenizers 0.12.1