metadata
tags:
- generated_from_trainer
metrics:
- wer
- cer
model-index:
- name: wav2vec2-large-asr-th-2
results: []
datasets:
- common_voice
- mozilla-foundation/common_voice_10_0
language:
- th
pipeline_tag: automatic-speech-recognition
library_name: transformers
wav2vec2-large-asr-th-2
This model was find-tune from on the CommonVoice dataset. It achieves the following results on the evaluation set:
- Loss: 0.2310
- Wer: 32.99%
- Cer: 3.75%
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: 12
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 36
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.065 | 0.18 | 1000 | 0.5433 | 0.3259 | 0.0891 |
| 0.0792 | 0.36 | 2000 | 0.5453 | 0.3269 | 0.0901 |
| 0.1663 | 0.53 | 3000 | 0.4702 | 0.3299 | 0.0908 |
| 0.7971 | 0.71 | 4000 | 0.2513 | 0.3244 | 0.0889 |
| 0.7588 | 0.89 | 5000 | 0.2310 | 0.3196 | 0.0878 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3