Whisper small Hi - CKS 2511 finalbatch
This model is a fine-tuned version of CKSINGH/whisper-small-hi-graminVoice on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5777
- Wer: 33.1148
- Cer: 14.1232
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: 1.75e-05
- train_batch_size: 16
- 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: 500
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.1067 | 6.94 | 500 | 0.4508 | 36.7213 | 17.2021 |
| 0.0083 | 13.89 | 1000 | 0.5445 | 35.4098 | 15.9973 |
| 0.001 | 20.83 | 1500 | 0.5720 | 34.7541 | 14.9264 |
| 0.0001 | 27.78 | 2000 | 0.5777 | 33.1148 | 14.1232 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
- Downloads last month
- 8