Whisper Large V3 Taiwanese Hakka

🚀 whisper-large-v3-taiwanese-hakka

This model is a fine - tuned version of the Taiwanese Hakka openai/whisper-large-v3. It uses dialect IDs as prompts during training to test if adding prompts in multi - dialect fine - tuning of Whisper yields better results.

🚀 Quick Start

This model is a fine - tuned version of the Taiwanese Hakka openai/whisper-large-v3. During training, it uses the IDs of each dialect as prompts to experiment whether adding prompts to the fine - tuning of Whisper when using multiple dialects will lead to better results.

✨ Features

Dialect and Id

• Four - County: htia_sixian
• Hailu: htia_hailu
• Dapu: htia_dapu
• Raoping: htia_raoping
• Zhaoan: htia_zhaoan
• Southern Four - County: htia_nansixian

Training process

The model was trained with the following hyperparameters:

• Batch size: 32
• Epochs: 3
• Warmup Steps: 50
• Total Steps: 42549
• Learning rate: 7e - 5
• Data augmentation: No

📦 Installation

No specific installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

Basic Usage

import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
device = "cuda:0" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
model_id = "formospeech/whisper-large-v3-taiwanese-hakka"
dialect_id = "htia_sixian"
model = AutoModelForSpeechSeq2Seq.from_pretrained(
    model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
)
model.to(device)
processor = AutoProcessor.from_pretrained(model_id)
pipe = pipeline(
    "automatic-speech-recognition",
    model=model,
    tokenizer=processor.tokenizer,
    feature_extractor=processor.feature_extractor,
    max_new_tokens=128,
    chunk_length_s=30,
    batch_size=16,
    torch_dtype=torch_dtype,
    device=device,
)
generate_kwargs = {"language": "Chinese", "prompt_ids": torch.from_numpy(processor.get_prompt_ids(dialect_id)).to(device)}
transcription = pipe("path/to/my_audio.wav", generate_kwargs=generate_kwargs)
print(transcription.replace(f" {dialect_id}", ""))

Advanced Usage

There is no advanced usage example in the original document, so this part is skipped.

📚 Documentation

There is no more detailed documentation content in the original document, so this section is skipped.

🔧 Technical Details

There is no in - depth technical details content (more than 50 words) in the original document, so this section is skipped.

📄 License