mms-1b-all-nyagen-combined-train-drodat-gdro-2.5e-3-dat-0.0012-no-sd-0.00-52
This model is a fine-tuned version of facebook/mms-1b-all on the NYAGEN - NYA dataset. It achieves the following results on the evaluation set:
- Loss: 37.8789
- Cer: 0.0836
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: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 430.873 | 1.2566 | 500 | 415.6341 | 0.8675 |
| 295.1986 | 2.5132 | 1000 | 282.9854 | 0.9635 |
| 28.5728 | 3.7698 | 1500 | 48.1279 | 0.0988 |
| 22.453 | 5.0252 | 2000 | 42.9916 | 0.0912 |
| 21.738 | 6.2818 | 2500 | 39.8314 | 0.0873 |
| 17.6986 | 7.5384 | 3000 | 40.8975 | 0.0846 |
| 19.7477 | 8.7950 | 3500 | 38.8177 | 0.0832 |
| 16.8082 | 10.0503 | 4000 | 38.5422 | 0.0830 |
| 18.0795 | 11.3069 | 4500 | 37.8784 | 0.0836 |
| 18.2335 | 12.5635 | 5000 | 40.1897 | 0.0817 |
| 13.331 | 13.8201 | 5500 | 39.8742 | 0.0816 |
| 16.8736 | 15.0755 | 6000 | 39.8727 | 0.0810 |
Framework versions
- Transformers 4.53.0
- Pytorch 2.10.0+cu128
- Datasets 4.5.0
- Tokenizers 0.21.4