metadata
language:
- sq
license: apache-2.0
library_name: transformers
tags:
- generated_from_trainer
base_model: openai/whisper-medium
datasets:
- Kushtrim/common_voice_19_sq
metrics:
- wer
model-index:
- name: Whisper Medium SQ
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: Common Voice 19.0
type: Kushtrim/common_voice_19_sq
args: 'config: sq, split: test'
metrics:
- type: wer
value: 7.322033898305085
name: Wer
Whisper Medium SQ
This model is a fine-tuned version of openai/whisper-medium on the Common Voice 19.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0979
- Wer: 7.3220
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.5161 | 0.4237 | 250 | 0.4949 | 38.8136 |
| 0.3207 | 0.8475 | 500 | 0.3084 | 29.4689 |
| 0.1595 | 1.2712 | 750 | 0.2255 | 21.6949 |
| 0.1239 | 1.6949 | 1000 | 0.1733 | 16.4859 |
| 0.0488 | 2.1186 | 1250 | 0.1338 | 13.0847 |
| 0.04 | 2.5424 | 1500 | 0.1188 | 10.8136 |
| 0.0241 | 2.9661 | 1750 | 0.1023 | 8.7684 |
| 0.0075 | 3.3898 | 2000 | 0.1022 | 8.1695 |
| 0.0065 | 3.8136 | 2250 | 0.0979 | 7.3220 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.0+cu121
- Datasets 3.0.0
- Tokenizers 0.19.1