metadata
library_name: peft
language:
- en
license: mit
base_model: openai/whisper-large-v3-turbo
tags:
- wft
- whisper
- automatic-speech-recognition
- audio
- speech
- generated_from_trainer
datasets:
- ntnu-smil/lttc-rebalanced-1-split
metrics:
- wer
model-index:
- name: whisper-large-v3-turbo-score-5-rebalanced-1
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: ntnu-smil/lttc-rebalanced-1-split
type: ntnu-smil/lttc-rebalanced-1-split
metrics:
- type: wer
value: 39.732142857142854
name: Wer
whisper-large-v3-turbo-score-5-rebalanced-1
This model is a fine-tuned version of openai/whisper-large-v3-turbo on the ntnu-smil/lttc-rebalanced-1-split dataset. It achieves the following results on the evaluation set:
- Loss: 3.9922
- Wer: 39.7321
- Cer: 25.9187
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: 0.0005
- train_batch_size: 4
- 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
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.0378 | 1.0 | 18 | 3.6518 | 40.3159 | 26.1279 |
| 0.0389 | 2.0 | 36 | 3.8285 | 40.0412 | 26.6444 |
| 0.0023 | 3.0 | 54 | 4.0319 | 40.4876 | 26.5529 |
| 0.0021 | 4.0 | 72 | 3.9976 | 39.3544 | 25.5656 |
| 0.0004 | 5.0 | 90 | 3.9922 | 39.7321 | 25.9187 |
Framework versions
- PEFT 0.13.2
- Transformers 4.46.3
- Pytorch 2.2.0+cu121
- Datasets 3.1.0
- Tokenizers 0.20.3