zwan074/maori_ASR

Usage

Whisper large-v3 is supported in Hugging Face 🤗 Transformers through the main branch in the Transformers repo. To run the model, first install the Transformers library through the GitHub repo.

pip install --upgrade pip
pip install --upgrade git+https://github.com/huggingface/transformers.git accelerate datasets[audio]

The model can be used with the pipeline class to transcribe audio files of arbitrary length. Transformers uses a chunked algorithm to transcribe long-form audio files, which in-practice is 9x faster than the sequential algorithm proposed by OpenAI (see Table 7 of the Distil-Whisper paper). The batch size should be set based on the specifications of your device:

import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
from datasets import load_dataset


device = "cuda:0" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32

model_id = "zwan074/maori_ASR"

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,
    return_timestamps=True,
    torch_dtype=torch_dtype,
    device=device,
)

To transcribe a local audio file, simply pass the path to your audio file when you call the pipeline:

result = pipe("audio.mp3")

print(result["text"])

Whisper predicts the language of the source audio automatically. If the source audio language is known a-priori, it can be passed as an argument to the pipeline:

result = pipe(sample, generate_kwargs={"language": "maori"})