<!-- 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. -->
kkkh_whisper_small_distillation_att_loss_libri360_epochs_50_batch_4_concat_dataset_try6
This model is a fine-tuned version of rohitp1/kkkh_whisper_small_distillation_att_loss_libri360_epochs_100_batch_4_concat_dataset on the None dataset. It achieves the following results on the evaluation set:
- Loss: 4.7728
- Wer: 4.5237
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: 8
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 256
- total_train_batch_size: 2048
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine_with_restarts
- lr_scheduler_warmup_ratio: 0.2
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.0773 | 0.98 | 100 | 4.9642 | 4.4263 |
| 6.1998 | 1.97 | 200 | 5.0002 | 4.3767 |
| 6.298 | 2.95 | 300 | 5.1708 | 4.3390 |
| 6.4775 | 3.94 | 400 | 5.3493 | 4.3491 |
| 6.5852 | 4.92 | 500 | 5.1604 | 4.3813 |
| 6.3406 | 5.91 | 600 | 4.9918 | 4.6671 |
| 6.6247 | 6.89 | 700 | 5.5665 | 4.4796 |
| 6.0813 | 7.88 | 800 | 4.7728 | 4.5237 |
Framework versions
- Transformers 4.27.3
- Pytorch 1.12.1
- Datasets 2.6.1
- Tokenizers 0.13.1
