app.py

import os

import gradio as gr
import librosa
import spaces
import torch
from huggingface_hub import login
from transformers import (
    WhisperFeatureExtractor,
    WhisperForConditionalGeneration,
    WhisperTokenizerFast,
)


_orig_init = WhisperTokenizerFast.__init__


def _patched_init(self, *args, **kwargs):
    est = kwargs.get("extra_special_tokens")
    if isinstance(est, list):
        kwargs["extra_special_tokens"] = {t: t for t in est}
    return _orig_init(self, *args, **kwargs)


WhisperTokenizerFast.__init__ = _patched_init


HF_TOKEN = os.getenv("HF_TOKEN")
if HF_TOKEN:
    login(token=HF_TOKEN)

MODEL_REPO = os.getenv("MODEL_REPO_ID", "Batuka0901/MN_ASR")
SAMPLING_RATE = 16000

print(f"Loading {MODEL_REPO} (CPU at startup) ...")
model = WhisperForConditionalGeneration.from_pretrained(MODEL_REPO, token=HF_TOKEN)
model.eval()
tokenizer = WhisperTokenizerFast.from_pretrained(MODEL_REPO, token=HF_TOKEN)
feature_extractor = WhisperFeatureExtractor.from_pretrained(MODEL_REPO, token=HF_TOKEN)
print("Model loaded.")
_moved_to_cuda = False


WAITING = "Төлөв: **Аудио хүлээж байна...**"
READY = "Төлөв: **Илгээхэд бэлэн.**"
WORKING = "Төлөв: **Танилт хийж байна...**"
DONE = "Төлөв: **Дууссан.**"


@spaces.GPU(duration=60)
def transcribe(audio_path):
    global _moved_to_cuda
    if not audio_path:
        return "", WAITING
    try:
        if not _moved_to_cuda and torch.cuda.is_available():
            model.to("cuda")
            _moved_to_cuda = True
        device = "cuda" if (_moved_to_cuda and torch.cuda.is_available()) else "cpu"

        audio, _ = librosa.load(audio_path, sr=SAMPLING_RATE)
        inputs = feature_extractor(
            audio, sampling_rate=SAMPLING_RATE, return_tensors="pt"
        )
        input_features = inputs.input_features.to(device)

        with torch.no_grad():
            predicted_ids = model.generate(
                input_features, language="mn", task="transcribe"
            )
        text = tokenizer.batch_decode(predicted_ids, skip_special_tokens=True)[0]
        return text.strip(), DONE
    except Exception as e:
        return "", f"Төлөв: **Алдаа** — {type(e).__name__}: {e}"


def on_audio_change(audio_path):
    if audio_path:
        return gr.update(interactive=True), READY
    return gr.update(interactive=False), WAITING


def on_clear():
    return gr.update(interactive=False), WAITING


INSTRUCTIONS = """
### Заавар

1. **Audio оруулна уу** — файл upload хийх эсвэл микрофоноор шууд бичлэг хийнэ
2. **Generate** товчийг дарна — таны хэлсэн үгийг загвар таньж текст болгоно

"""

CSS = """
footer { display: none !important; visibility: hidden !important; }
.gradio-container > .footer { display: none !important; }
button.api-link, .api-docs, a[href*="/api/"] { display: none !important; }
"""

with gr.Blocks(title="Speech to Text", css=CSS) as demo:
    with gr.Tab("Speech to Text"):
        gr.Markdown(INSTRUCTIONS)
        with gr.Row():
            with gr.Column(scale=1):
                audio_in = gr.Audio(
                    sources=["upload", "microphone"],
                    type="filepath",
                    label="Audio",
                )
                with gr.Row():
                    clear_btn = gr.Button("Clear", variant="secondary", size="sm")
                    submit_btn = gr.Button(
                        "Generate", variant="primary", size="sm", interactive=False
                    )
            with gr.Column(scale=1):
                text_out = gr.Textbox(label="Гаралт", lines=10)

        status = gr.Markdown(WAITING)

        audio_in.change(on_audio_change, inputs=audio_in, outputs=[submit_btn, status])
        submit_btn.click(lambda: WORKING, outputs=status).then(
            transcribe, inputs=audio_in, outputs=[text_out, status]
        )
        clear_btn.click(
            lambda: (None, "", WAITING),
            outputs=[audio_in, text_out, status],
        ).then(on_clear, outputs=[submit_btn, status])


if __name__ == "__main__":
    demo.queue().launch(show_api=False)