<!-- 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. -->
dgx1_whisper_tiny_teacher_student_distillation_mozilla_epochs_100_batch_32_concatenate_dataset
This model is a fine-tuned version of rohitp1/dgx1_whisper_tiny_finetune_teacher_babble_noise_mozilla_40_epochs_batch_32 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.6990
- Wer: 35.5686
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: 32
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 128
- total_train_batch_size: 4096
- 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: 100
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.6059 | 14.7 | 500 | 1.6814 | 34.3395 |
| 0.3409 | 29.41 | 1000 | 1.7039 | 34.8113 |
| 0.213 | 44.12 | 1500 | 1.6941 | 35.2691 |
| 0.1475 | 58.82 | 2000 | 1.6869 | 35.4972 |
| 0.1135 | 73.53 | 2500 | 1.7040 | 35.5564 |
| 0.095 | 88.23 | 3000 | 1.6990 | 35.5686 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.12.1
- Datasets 2.8.0
- Tokenizers 0.13.2
