Update app.py

app.py

@@ -1,4 +1,5 @@
 import torch
+import torchaudio  # ✅ Added torchaudio to handle audio resampling
 import gradio as gr
 import time  
 import numpy as np
@@ -27,6 +28,7 @@ pipe = pipeline(
     chunk_length_s=2,  # ✅ Process in 2-second chunks for ultra-low latency
     torch_dtype=torch_dtype,
     device=device,
+    sampling_rate=16000,  # ✅ Explicitly set sampling rate to avoid resampling issues
 )
 
 # ✅ 4️⃣ Real-Time Streaming Transcription (Microphone)
@@ -42,84 +44,4 @@ def stream_transcribe(stream, new_chunk):
         y = y.astype(np.float32)
         y /= np.max(np.abs(y))
 
-        # ✅ Append to Stream
-        if stream is not None:
-            stream = np.concatenate([stream, y])
-        else:
-            stream = y
-            
-        # ✅ Run Transcription
-        transcription = pipe({"sampling_rate": sr, "raw": stream})["text"]
-        latency = time.time() - start_time
-
-        return stream, transcription, f"{latency:.2f} sec"
-    
-    except Exception as e:
-        print(f"Error: {e}")
-        return stream, str(e), "Error"
-
-# ✅ 5️⃣ Transcription for File Upload
-def transcribe(inputs, previous_transcription):
-    start_time = time.time()  
-    try:
-        # ✅ Convert file input to correct format
-        sample_rate, audio_data = inputs
-        transcription = pipe({"sampling_rate": sample_rate, "raw": audio_data})["text"]
-        
-        previous_transcription += transcription
-        latency = time.time() - start_time
-        
-        return previous_transcription, f"{latency:.2f} sec"
-
-    except Exception as e:
-        print(f"Error: {e}")
-        return previous_transcription, "Error"
-
-# ✅ 6️⃣ Clear Function
-def clear():
-    return ""
-
-# ✅ 7️⃣ Gradio Interface (Microphone Streaming)
-with gr.Blocks() as microphone:
-    gr.Markdown(f"# Whisper Tiny - Real-Time Transcription (CPU) 🎙️")
-    gr.Markdown(f"Using [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) for ultra-fast speech-to-text.")
-
-    with gr.Row():
-        input_audio_microphone = gr.Audio(sources=["microphone"], type="numpy", streaming=True)
-        output = gr.Textbox(label="Live Transcription", value="")
-        latency_textbox = gr.Textbox(label="Latency (seconds)", value="0.0")
-
-    with gr.Row():
-        clear_button = gr.Button("Clear Output")
-
-    state = gr.State()
-    input_audio_microphone.stream(
-        stream_transcribe, [state, input_audio_microphone], 
-        [state, output, latency_textbox], time_limit=30, stream_every=1
-    )
-    clear_button.click(clear, outputs=[output])
-
-# ✅ 8️⃣ Gradio Interface (File Upload)
-with gr.Blocks() as file:
-    gr.Markdown(f"# Upload Audio File for Transcription 🎵")
-    gr.Markdown(f"Using [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) for speech-to-text.")
-
-    with gr.Row():
-        input_audio = gr.Audio(sources=["upload"], type="numpy")
-        output = gr.Textbox(label="Transcription", value="")
-        latency_textbox = gr.Textbox(label="Latency (seconds)", value="0.0")
-
-    with gr.Row():
-        submit_button = gr.Button("Submit")
-        clear_button = gr.Button("Clear Output")
-
-    submit_button.click(transcribe, [input_audio, output], [output, latency_textbox])
-    clear_button.click(clear, outputs=[output])
-
-# ✅ 9️⃣ Final Gradio App (Supports Microphone & File Upload)
-with gr.Blocks(theme=gr.themes.Ocean()) as demo:
-    gr.TabbedInterface([microphone, file], ["Microphone", "Upload Audio"])
-
-# ✅ 1️⃣0️⃣ Run Gradio Locally
-if __name__ == "__main__":
-    demo.launch()
+        # ✅ Resample audio