Whisper Small Hi - Sanchit Gandhi
This model is a fine-tuned version of nurzhanit/whisper-enhanced-ml on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0005
- Wer: 0.0516
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: 1e-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: 25
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.0929 | 1.0 | 50 | 0.0404 | 8.1114 |
0.0385 | 2.0 | 100 | 0.0161 | 3.0761 |
0.0176 | 3.0 | 150 | 0.0081 | 1.4779 |
0.0087 | 4.0 | 200 | 0.0034 | 0.6358 |
0.0038 | 5.0 | 250 | 0.0020 | 0.4296 |
0.0025 | 6.0 | 300 | 0.0011 | 0.1203 |
0.0017 | 7.0 | 350 | 0.0008 | 0.0516 |
0.0009 | 8.0 | 400 | 0.0006 | 0.0516 |
0.0007 | 9.0 | 450 | 0.0006 | 0.0516 |
0.0007 | 10.0 | 500 | 0.0005 | 0.0516 |
Framework versions
- Transformers 4.40.0
- Pytorch 2.5.0+cu124
- Datasets 3.0.2
- Tokenizers 0.19.1
- Downloads last month
- 0
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no library tag.
Model tree for nurzhanit/whisper-omg-2.5
Unable to build the model tree, the base model loops to the model itself. Learn more.