---
library_name: transformers
license: mit
datasets:
- AlienKevin/mixed_cantonese_and_english_speech
- mozilla-foundation/common_voice_17_0
- mozilla-foundation/common_voice_11_0
metrics:
- cer
base_model:
- openai/whisper-small
---

CER: 15.4% <br>

transformers-4.46.3

Train Args: <br>
per_device_train_batch_size=32, <br>
gradient_accumulation_steps=1,  <br>
learning_rate=1e-5, <br>
gradient_checkpointing=True, <br>
per_device_eval_batch_size=64, <br>
generation_max_length=225, <br>

Hardware: <br>
NVIDIA Tesla V100 16GB * 4 <br>

FAQ:
1. If having tokenizer issue during inference, please update your transformers version to >= 4.46.3

```bash
pip install --upgrade transformers==4.46.3
```