<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
large_finetune_M01
This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.9064
- Wer: 0.5667
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: 0.0001
- train_batch_size: 15
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 75
- num_epochs: 250
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.1813 | 25.0 | 150 | 1.6666 | 0.8667 |
| 1.0121 | 50.0 | 300 | 1.6864 | 0.7667 |
| 0.0471 | 75.0 | 450 | 1.1663 | 0.6333 |
| 0.0524 | 100.0 | 600 | 1.6732 | 0.7 |
| 0.0205 | 125.0 | 750 | 1.9236 | 0.5667 |
| 0.0156 | 150.0 | 900 | 1.8241 | 0.5 |
| 0.0069 | 175.0 | 1050 | 1.6829 | 0.5 |
| 0.007 | 200.0 | 1200 | 1.9494 | 0.6 |
| 0.0022 | 225.0 | 1350 | 2.0226 | 0.5667 |
| 0.0017 | 250.0 | 1500 | 1.9064 | 0.5667 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.2+cu102
- Datasets 2.3.2
- Tokenizers 0.12.1
