my_sanskrit_model
This model is a fine-tuned version of t5-small on the itihasa dataset. It achieves the following results on the evaluation set:
- Loss: 3.5101
- Bleu: 0.2607
- Gen Len: 18.9973
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: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 3.9557 | 1.0 | 4698 | 3.7191 | 0.3291 | 18.9973 |
| 3.8243 | 2.0 | 9396 | 3.6068 | 0.2728 | 18.9973 |
| 3.7562 | 3.0 | 14094 | 3.5503 | 0.2911 | 18.9973 |
| 3.7306 | 4.0 | 18792 | 3.5207 | 0.2404 | 18.9973 |
| 3.7003 | 5.0 | 23490 | 3.5101 | 0.2607 | 18.9973 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2