Configuration Parsing
Warning:
In adapter_config.json: "peft.task_type" must be a string
Whisper medium Fa Yazdi - SRezaS
This model is a fine-tuned version of openai/whisper-medium on the Persian Yazdi Accent Custom Dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.2112
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: 0.0001
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use 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: 50
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 1.0 | 16 | 0.6912 |
| No log | 2.0 | 32 | 0.4918 |
| No log | 3.0 | 48 | 0.3308 |
| 0.6073 | 4.0 | 64 | 0.2547 |
| 0.6073 | 5.0 | 80 | 0.2162 |
| 0.6073 | 6.0 | 96 | 0.2000 |
| 0.1452 | 7.0 | 112 | 0.1921 |
| 0.1452 | 8.0 | 128 | 0.1982 |
| 0.1452 | 9.0 | 144 | 0.1955 |
| 0.0408 | 10.0 | 160 | 0.1991 |
| 0.0408 | 11.0 | 176 | 0.1980 |
| 0.0408 | 12.0 | 192 | 0.2016 |
| 0.015 | 13.0 | 208 | 0.2068 |
| 0.015 | 14.0 | 224 | 0.2026 |
| 0.015 | 15.0 | 240 | 0.2052 |
| 0.0082 | 16.0 | 256 | 0.2086 |
| 0.0082 | 17.0 | 272 | 0.2061 |
| 0.0082 | 18.0 | 288 | 0.2095 |
| 0.0056 | 19.0 | 304 | 0.2078 |
| 0.0056 | 20.0 | 320 | 0.2112 |
Framework versions
- PEFT 0.13.3.dev0
- Transformers 4.47.0
- Pytorch 2.5.1+cu124
- Datasets 3.1.0
- Tokenizers 0.21.0
- Downloads last month
- 66