mms-1b-all-bemgen-combined-train-drodat-gdro-3e-2-dat-0.00-no-sd-0.00-42
This model is a fine-tuned version of facebook/mms-1b-all on the BEMGEN - BEM dataset. It achieves the following results on the evaluation set:
- Loss: 12.7139
- Cer: 0.0530
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: 8
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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 |
|---|---|---|---|---|
| 277.5065 | 1.2690 | 500 | 233.8936 | 0.9717 |
| 35.791 | 2.5381 | 1000 | 14.4497 | 0.0629 |
| 31.0339 | 3.8071 | 1500 | 13.8997 | 0.0612 |
| 30.5563 | 5.0761 | 2000 | 13.5236 | 0.0582 |
| 30.2489 | 6.3452 | 2500 | 13.2939 | 0.0580 |
| 27.5266 | 7.6142 | 3000 | 13.2561 | 0.0561 |
| 25.2992 | 8.8832 | 3500 | 13.0463 | 0.0563 |
| 27.1152 | 10.1523 | 4000 | 12.9897 | 0.0565 |
| 24.5777 | 11.4213 | 4500 | 12.7654 | 0.0549 |
| 25.2812 | 12.6904 | 5000 | 12.8302 | 0.0538 |
| 23.6094 | 13.9594 | 5500 | 12.7142 | 0.0530 |
| 22.5199 | 15.2284 | 6000 | 12.9666 | 0.0535 |
| 23.2766 | 16.4975 | 6500 | 12.9791 | 0.0531 |
| 24.5292 | 17.7665 | 7000 | 12.7273 | 0.0539 |
Framework versions
- Transformers 4.53.0
- Pytorch 2.10.0+cu128
- Datasets 4.5.0
- Tokenizers 0.21.4