metadata
library_name: transformers
language:
- ro
license: apache-2.0
base_model: openai/whisper-base
tags:
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_17_0
metrics:
- wer
model-index:
- name: Whisper Base Ro - Augustin Jianu
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 17.0
type: mozilla-foundation/common_voice_17_0
config: ro
split: None
args: 'config: ro, split: test'
metrics:
- name: Wer
type: wer
value: 31.09085081377542
Whisper Base Ro - Augustin Jianu
This model is a fine-tuned version of openai/whisper-base on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4626
- Wer: 31.0909
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: 5e-06
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- 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: 500
- training_steps: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.366 | 1.7730 | 1000 | 0.4236 | 35.2256 |
| 0.1676 | 3.5461 | 2000 | 0.3700 | 31.5503 |
| 0.0752 | 5.3191 | 3000 | 0.3683 | 30.3287 |
| 0.0355 | 7.0922 | 4000 | 0.3841 | 30.1756 |
| 0.025 | 8.8652 | 5000 | 0.4003 | 30.0011 |
| 0.0106 | 10.6383 | 6000 | 0.4232 | 31.6820 |
| 0.0067 | 12.4113 | 7000 | 0.4380 | 31.4221 |
| 0.0043 | 14.1844 | 8000 | 0.4520 | 30.1613 |
| 0.0038 | 15.9574 | 9000 | 0.4594 | 30.1079 |
| 0.0032 | 17.7305 | 10000 | 0.4626 | 31.0909 |
Framework versions
- Transformers 4.48.0
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0