metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-tiny-hindi2_test
results: []
whisper-tiny-hindi2_test
This model is a fine-tuned version of openai/whisper-tiny on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1752
- Wer: 27.1501
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: 3.75e-05
- train_batch_size: 16
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 50
- training_steps: 200
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2566 | 1.2698 | 40 | 0.1388 | 23.6931 |
| 0.0988 | 2.5397 | 80 | 0.1452 | 23.4401 |
| 0.0499 | 3.8095 | 120 | 0.1461 | 23.9460 |
| 0.0299 | 5.0794 | 160 | 0.1596 | 24.1990 |
| 0.0132 | 6.3492 | 200 | 0.1752 | 27.1501 |
Framework versions
- Transformers 4.41.1
- Pytorch 2.3.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1