metadata
language:
- sq
license: apache-2.0
base_model: openai/whisper-small
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- Kaggle
metrics:
- wer
model-index:
- name: Whisper tiny Albanian Test
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Kaggle Albanian
type: Kaggle
args: 'config: sq, split: test'
metrics:
- name: Wer
type: wer
value: 35.183525075949994
Whisper tiny Albanian Test
This model is a fine-tuned version of openai/whisper-small on the Kaggle Albanian dataset. It achieves the following results on the evaluation set:
- Loss: 0.4063
- Wer: 35.1835
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: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 3000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.5598 | 0.24 | 500 | 0.6847 | 55.6153 |
| 0.4449 | 0.49 | 1000 | 0.5154 | 44.8005 |
| 0.3986 | 0.73 | 1500 | 0.4648 | 40.5448 |
| 0.3684 | 0.97 | 2000 | 0.4242 | 37.3848 |
| 0.2517 | 1.22 | 2500 | 0.4143 | 35.9306 |
| 0.2467 | 1.46 | 3000 | 0.4063 | 35.1835 |
Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2