metadata
library_name: transformers
language:
- th
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_17_0
metrics:
- wer
model-index:
- name: Whisper Small Th Combined Finetuned
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 17.0
type: mozilla-foundation/common_voice_17_0
config: th
split: test
args: 'config: th, split: validated'
metrics:
- name: Wer
type: wer
value: 0.4430396682052311
Whisper Small Th Combined Finetuned
This model is a fine-tuned version of openai/whisper-small on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0858
- Wer: 0.4430
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: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- total_eval_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 8000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2913 | 0.2175 | 1000 | 0.2658 | 0.7758 |
| 0.2112 | 0.4349 | 2000 | 0.1918 | 0.6780 |
| 0.1733 | 0.6524 | 3000 | 0.1544 | 0.6206 |
| 0.1485 | 0.8698 | 4000 | 0.1279 | 0.5651 |
| 0.1029 | 1.0873 | 5000 | 0.1102 | 0.5119 |
| 0.0989 | 1.3047 | 6000 | 0.0983 | 0.4775 |
| 0.0935 | 1.5222 | 7000 | 0.0901 | 0.4566 |
| 0.0863 | 1.7396 | 8000 | 0.0858 | 0.4430 |
Framework versions
- Transformers 4.45.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0