metadata
language:
- zh
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
datasets:
- formospeech/hat_asr_aligned
model-index:
- name: Whisper Tiny Hakka Condenser
results: []
Whisper Tiny Hakka Condenser
This model is a fine-tuned version of openai/whisper-tiny on the HAT ASR Aligned dataset. It achieves the following results on the evaluation set:
- Loss: 0.2216
- Cer: 13.1863
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: 64
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 976
- training_steps: 9760
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 1.2175 | 0.9980 | 488 | 1.2419 | 50.2601 |
| 0.3915 | 1.9959 | 976 | 0.5156 | 27.2673 |
| 0.1993 | 2.9939 | 1464 | 0.3351 | 18.1346 |
| 0.121 | 3.9918 | 1952 | 0.2783 | 16.5268 |
| 0.0808 | 4.9898 | 2440 | 0.2555 | 15.1964 |
| 0.0538 | 5.9877 | 2928 | 0.2460 | 14.7722 |
| 0.0348 | 6.9857 | 3416 | 0.2305 | 14.2647 |
| 0.0255 | 7.9836 | 3904 | 0.2224 | 13.6105 |
| 0.019 | 8.9816 | 4392 | 0.2232 | 14.8635 |
| 0.0126 | 9.9796 | 4880 | 0.2214 | 13.4857 |
| 0.0079 | 10.9775 | 5368 | 0.2234 | 13.6510 |
| 0.0058 | 11.9755 | 5856 | 0.2211 | 13.5261 |
| 0.0045 | 12.9734 | 6344 | 0.2206 | 13.9920 |
| 0.0034 | 13.9714 | 6832 | 0.2210 | 13.8082 |
| 0.0029 | 14.9693 | 7320 | 0.2235 | 12.1090 |
| 0.0025 | 15.9673 | 7808 | 0.2203 | 12.2974 |
| 0.0022 | 16.9652 | 8296 | 0.2217 | 12.2847 |
| 0.002 | 17.9632 | 8784 | 0.2218 | 13.2291 |
| 0.0018 | 18.9611 | 9272 | 0.2216 | 13.3285 |
| 0.0018 | 19.9591 | 9760 | 0.2216 | 13.1863 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1