bart-hindi
This model is a fine-tuned version of facebook/bart-base on the Someman/hindi-summarization dataset. It achieves the following results on the evaluation set:
- Loss: 0.4985
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: 5e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.6568 | 0.14 | 500 | 0.6501 |
| 0.682 | 0.29 | 1000 | 0.5757 |
| 0.5331 | 0.43 | 1500 | 0.5530 |
| 0.5612 | 0.58 | 2000 | 0.5311 |
| 0.5685 | 0.72 | 2500 | 0.5043 |
| 0.4993 | 0.87 | 3000 | 0.4985 |
Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3