Update app.py

app.py

@@ -17,47 +17,28 @@ demo = gr.Interface(
 
 demo.launch()
 
-
 from pytube import YouTube
+from moviepy.editor import VideoFileClip
+from transformers import WhisperProcessor, WhisperForConditionalGeneration
+import librosa
 
+# Step 1: Download YouTube video as audio
 video_url = "https://www.youtube.com/watch?v=YOUR_VIDEO_ID"
 yt = YouTube(video_url)
 stream = yt.streams.filter(only_audio=True).first()
 stream.download(filename="video_audio.mp4")
 
-from moviepy.editor import VideoFileClip
-
+# Step 2: Extract audio as WAV
 video = VideoFileClip("video_audio.mp4")
 audio = video.audio
 audio.write_audiofile("output_audio.wav")
 
-
-
-from transformers import WhisperProcessor, WhisperForConditionalGeneration
-import torch
-from datasets import load_dataset
-
-# Load model and processor
+# Step 3: Speech-to-text with Whisper-Small
 processor = WhisperProcessor.from_pretrained("openai/whisper-small")
 model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small")
-
-# Optional: Use GPU if available
-device = "cuda" if torch.cuda.is_available() else "cpu"
-model = model.to(device)
-
-# Load sample audio (here using a dummy dataset, aap apni audio file bhi use kar sakte hain)
-ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
-sample = ds[0]["audio"]
-
-# Prepare audio input
-input_features = processor(sample["array"], sampling_rate=sample["sampling_rate"], return_tensors="pt").input_features
-input_features = input_features.to(device)
-
-# Generate transcription
+audio, sr = librosa.load("output_audio.wav", sr=16000)
+input_features = processor(audio, sampling_rate=sr, return_tensors="pt").input_features
 predicted_ids = model.generate(input_features)
-
-# Decode transcription
-transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
+transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
 print(transcription)
 
-