TrorYongASR2-tiny - test
This model is a fine-tuned version of KrorngAI/TrorYongASR2-tiny on the Khmer MPWT Speech dataset. It achieves the following results on the evaluation set:
- Loss: 6.2215
- Wer: 43.1121
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 40
- training_steps: 400
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0 | 0 | 13.6963 | 69.6598 |
| 8.9500 | 0.8621 | 100 | 8.1658 | 75.8505 |
| 7.5705 | 1.7241 | 200 | 6.6443 | 50.0837 |
| 6.4357 | 2.5862 | 300 | 6.3206 | 43.7256 |
| 6.5587 | 3.4483 | 400 | 6.2215 | 43.1121 |
Framework versions
- Transformers 5.20.1
- Pytorch 2.11.0+cu130
- Datasets 4.8.5
- Tokenizers 0.23.3
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for Kimang18/TrorYongASR2-tiny-km
Base model
KrorngAI/TrorYongASR2-tinyDataset used to train Kimang18/TrorYongASR2-tiny-km
Evaluation results
- Wer on Khmer MPWT Speechself-reported43.112