Whisper Tiny MNSC ASR4
This model is a fine-tuned version of openai/whisper-tiny on the Multitask National Speech Corpus v1 (ASR-PART4) dataset. It achieves the following results on the evaluation set:
- Loss: 1.9557
- Wer: 96.1757
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
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 25
- training_steps: 125
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.7344 | 0.3987 | 60 | 2.0517 | 98.4314 |
| 1.4443 | 0.7973 | 120 | 1.9557 | 96.1757 |
Framework versions
- Transformers 4.51.1
- Pytorch 2.6.0+cu124
- Datasets 3.5.0
- Tokenizers 0.21.1
- Downloads last month
- 77
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
Model tree for urroxyz/whisper-tiny-mnsc-asr4
Base model
openai/whisper-tinyDataset used to train urroxyz/whisper-tiny-mnsc-asr4
Evaluation results
- Wer on Multitask National Speech Corpus v1 (ASR-PART4)self-reported96.176