mbart-large-cc25-finetuned-hi-to-en-v1
This model is a fine-tuned version of facebook/mbart-large-cc25 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.4978
- Bleu: 33.3366
- Gen Len: 22.7806
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: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|---|---|---|---|---|---|
| 1.6774 | 1.0 | 3955 | 1.5499 | 7.9551 | 73.7518 |
| 1.2296 | 2.0 | 7910 | 1.4846 | 32.8075 | 23.7341 |
| 0.9127 | 3.0 | 11865 | 1.5345 | 31.9747 | 23.6264 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 2.0.0
- Tokenizers 0.11.6