Update app.py

app.py

@@ -1,44 +1,46 @@
-with open("app.py", "w") as f:
-    f.write("""
+import os
 import gradio as gr
+import torch
 from transformers import pipeline
-import os
 
-print("🎤 Loading transcription pipeline...")
-try:
-    transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-small", device="cuda:0")
-except Exception as e:
-    print(f"Could not load model on GPU: {e}. Trying on CPU.")
-    transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-small", device="cpu")
+MODEL_ID = "openai/whisper-small"
+
+def load_asr():
+    # Prefer GPU if available, else CPU. For transformers pipelines:
+    # device: int index for CUDA, or -1 for CPU.
+    device = 0 if torch.cuda.is_available() else -1
+    print(f"🎤 Loading transcription pipeline on {'GPU' if device == 0 else 'CPU'}...")
+    return pipeline(
+        task="automatic-speech-recognition",
+        model=MODEL_ID,
+        device=device
+    )
 
+asr = load_asr()
 
 def transcribe_audio(audio_file_path):
-    if audio_file_path is None:
+    if not audio_file_path:
         return "Please upload an audio file."
-
     if not os.path.exists(audio_file_path):
-        return f"Error: Audio file not found at {audio_file_path}"
+        return f"Error: file not found at {audio_file_path}"
 
-    print(f" transcribe the audio file: {audio_file_path}")
+    print(f"→ Transcribing: {audio_file_path}")
     try:
-        transcription_result = transcriber(audio_file_path, chunk_length_s=30, return_timestamps=True)
-        return transcription_result["text"]
+        # chunk_length_s works with Whisper in transformers
+        result = asr(audio_file_path, chunk_length_s=30, return_timestamps=True)
+        # result is a dict with "text" and possibly "chunks"
+        return result.get("text", "").strip() or "(No text recognized)"
     except Exception as e:
         return f"Error during transcription: {e}"
 
-
-print("🚀 Creating Gradio interface for Transcription...")
 iface = gr.Interface(
     fn=transcribe_audio,
-    inputs=gr.Audio(type="filepath", label="Upload Audio File"),
+    inputs=gr.Audio(type="filepath", label="Upload audio (MP3/WAV)"),
     outputs=gr.Textbox(label="Transcription"),
     title="Audio Transcription Pipeline",
-    description="Upload an audio file (e.g., MP3, WAV) to get a transcription."
+    description="Upload an audio file and get a Whisper-small transcription.",
 )
 
-print("✨ Launching Gradio interface...")
-iface.launch()
-print("\n✅ Gradio interface launched.")
-""")
-
-print("✅ Saved Gradio app code to app.py")
+if __name__ == "__main__":
+    # Bind to all interfaces for Docker/Spaces
+    iface.launch(server_name="0.0.0.0", server_port=7860)