metadata
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
datasets:
- common_voice_13_0
base_model: openai/whisper-tiny
model-index:
- name: arabic_speech_ASR_Whisper
results: []
arabic_speech_ASR_Whisper
This model is a fine-tuned version of openai/whisper-tiny on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.8332
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: 0.001
- train_batch_size: 6
- eval_batch_size: 6
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.9454 | 1.0 | 400 | 0.9238 |
| 0.846 | 2.0 | 800 | 0.8786 |
| 0.7349 | 3.0 | 1200 | 0.8355 |
| 0.6079 | 4.0 | 1600 | 0.8174 |
| 0.5083 | 5.0 | 2000 | 0.8075 |
| 0.4655 | 6.0 | 2400 | 0.8072 |
| 0.3587 | 7.0 | 2800 | 0.8087 |
| 0.3336 | 8.0 | 3200 | 0.8147 |
| 0.26 | 9.0 | 3600 | 0.8242 |
| 0.2193 | 10.0 | 4000 | 0.8332 |
Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2