Spaces:
Sleeping
Sleeping
File size: 4,015 Bytes
7f17e42 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 |
import os
import sys
import subprocess
import tempfile
import requests
from moviepy.editor import VideoFileClip
# Ensure the official OpenAI Whisper package is installed (supports load_model)
try:
import whisper
if not hasattr(whisper, 'load_model'):
raise ImportError
except ImportError:
subprocess.run([sys.executable, "-m", "pip", "install", "--upgrade", "openai-whisper"], check=True)
import whisper
import torch
import librosa
import pandas as pd
from transformers import Wav2Vec2Processor, Wav2Vec2ForSequenceClassification
from huggingface_hub import login
import gradio as gr
# Authenticate with Hugging Face (token via HF_TOKEN env var)
# Device setup (GPU if available)
device = 'cuda' if torch.cuda.is_available() else 'cpu'
def load_models():
# Load Whisper directly on the target device
whisper_model = whisper.load_model('base', device=device)
processor = Wav2Vec2Processor.from_pretrained(
'jonatasgrosman/wav2vec2-large-xlsr-53-english'
)
accent_model = Wav2Vec2ForSequenceClassification.from_pretrained(
'jonatasgrosman/wav2vec2-large-xlsr-53-english'
).to(device)
return whisper_model, processor, accent_model
whisper_model, processor, accent_model = load_models()
# Main analysis function
def analyze(video_url: str):
# Download video to temp file
with tempfile.NamedTemporaryFile(delete=False, suffix='.mp4') as tmp_vid:
response = requests.get(video_url, stream=True)
response.raise_for_status()
for chunk in response.iter_content(chunk_size=1024 * 1024):
if chunk:
tmp_vid.write(chunk)
video_path = tmp_vid.name
# Extract audio
audio_path = video_path.replace('.mp4', '.wav')
clip = VideoFileClip(video_path)
clip.audio.write_audiofile(audio_path, verbose=False, logger=None)
clip.close()
# Load audio waveform
speech, sr = librosa.load(audio_path, sr=16000)
# Transcribe with Whisper (model on correct device)
result = whisper_model.transcribe(speech)
transcript = result.get('text', '')
lang = result.get('language', 'unknown')
if lang != 'en':
transcript = f"[Non-English detected: {lang}]\n" + transcript
# Accent classification
inputs = processor(speech, sampling_rate=sr, return_tensors='pt', padding=True)
input_values = inputs.input_values.to(device)
attention_mask = inputs.attention_mask.to(device)
with torch.no_grad():
logits = accent_model(input_values=input_values, attention_mask=attention_mask).logits
probs = torch.softmax(logits, dim=-1).squeeze().cpu().tolist()
# Map default LABEL_x to human-readable accents
accent_labels = [
'American', 'Australian', 'British', 'Canadian', 'Indian',
'Irish', 'New Zealander', 'South African', 'Welsh'
] # ensure this matches model output order
accent_probs = [(accent_labels[i], probs[i] * 100) for i in range(len(probs))]
accent_probs.sort(key=lambda x: x[1], reverse=True)
top_accent, top_conf = accent_probs[0]
# Prepare DataFrame
df = pd.DataFrame(accent_probs, columns=['Accent', 'Confidence (%)'])
df = pd.DataFrame(accent_probs, columns=['Accent', 'Confidence (%)'])
# Cleanup temp files
try:
os.remove(video_path)
os.remove(audio_path)
except:
pass
return top_accent, f"{top_conf:.2f}%", df
# Gradio interface
interface = gr.Interface(
fn=analyze,
inputs=gr.Textbox(label='Video URL', placeholder='Enter public MP4 URL'),
outputs=[
# gr.Textbox(label='Transcript'),
gr.Textbox(label='Predicted Accent'),
gr.Textbox(label='Accent Confidence'),
gr.Dataframe(label='All Accent Probabilities')
],
title='English Accent Detector',
description='Paste a Loom or direct MP4 URL to extract, transcribe, and classify English accents (uses GPU if available).',
allow_flagging='never'
)
if __name__ == '__main__':
interface.launch()
|