Update app.py

app.py

@@ -1,100 +1,47 @@
-import torch
-from transformers import pipeline
 import gradio as gr
+from transformers import pipeline
 
-# Define the model details
-MODEL_NAME = "riteshkr/quantized-whisper-large-v3"  # Update with your actual model ID
-BATCH_SIZE = 8
-
-# Select device based on availability of CUDA (GPU) or fallback to CPU
-device = 0 if torch.cuda.is_available() else "cpu"
-
-# Load the ASR model pipeline
-pipe = pipeline(
-    task="automatic-speech-recognition",
-    model=MODEL_NAME,
-    chunk_length_s=30,
-    device=device,
-)
-
-# Utility function to format timestamps
-def format_timestamp(seconds: float, always_include_hours: bool = False, decimal_marker: str = "."):
-    if seconds is not None:
-        milliseconds = round(seconds * 1000.0)
-        hours = milliseconds // 3_600_000
-        milliseconds -= hours * 3_600_000
-        minutes = milliseconds // 60_000
-        milliseconds -= minutes * 60_000
-        seconds = milliseconds // 1_000
-        milliseconds -= seconds * 1_000
-        hours_marker = f"{hours:02d}:" if always_include_hours or hours > 0 else ""
-        return f"{hours_marker}{minutes:02d}:{seconds:02d}{decimal_marker}{milliseconds:03d}"
-    else:
-        return seconds
-
-# Transcription function for batch processing
-def transcribe(files, task, return_timestamps):
-    transcriptions = []
-    for file in files:  # Process each file in the batch
-        outputs = pipe(file, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=return_timestamps)
-        text = outputs["text"]
-        if return_timestamps:
-            timestamps = outputs["chunks"]
-            formatted_chunks = [
-                f"[{format_timestamp(chunk['timestamp'][0])} -> {format_timestamp(chunk['timestamp'][1])}] {chunk['text']}"
-                for chunk in timestamps
-            ]
-            text = "\n".join(formatted_chunks)
-        transcriptions.append(text)
-    return "\n\n".join(transcriptions)  # Return all transcriptions combined
+# Load the ASR model using the Hugging Face pipeline
+model_id = "riteshkr/quantized-whisper-large-v3"  # Update with your model path or ID
+pipe = pipeline("automatic-speech-recognition", model=model_id)
+
+# Define the transcription function
+def transcribe_speech(filepath):
+    output = pipe(
+        filepath,
+        max_new_tokens=256,
+        generate_kwargs={
+            "task": "transcribe",
+            "language": "english",
+        },  # Update the language as per your model's fine-tuning
+        chunk_length_s=30,
+        batch_size=8,
+    )
+    return output["text"]
 
-# Define Gradio interface for microphone input
+# Define the Gradio interface for microphone input
 mic_transcribe = gr.Interface(
-    fn=transcribe,
-    inputs=[
-        gr.Audio(sources="microphone", type="filepath"),
-        gr.Radio(["transcribe", "translate"], label="Task"),
-        gr.Checkbox(label="Return timestamps"),
-    ],
-    outputs="text",
-    layout="horizontal",
-    title="Whisper Demo: Transcribe Audio",
-    description=(
-        f"Transcribe long-form microphone inputs with the {MODEL_NAME} model. Supports transcription and translation."
-    ),
-    allow_flagging="never",
+    fn=transcribe_speech,
+    inputs=gr.Audio(sources="microphone", type="filepath"),
+    outputs=gr.Textbox(),
 )
 
-# Define Gradio interface for file upload
+# Define the Gradio interface for file upload input
 file_transcribe = gr.Interface(
-    fn=transcribe,
-    inputs=[
-        gr.Audio(sources="upload", type="filepath", label="Upload Audio File"),
-        gr.Radio(["transcribe", "translate"], label="Task"),
-        gr.Checkbox(label="Return timestamps"),
-    ],
-    outputs="text",
-    layout="horizontal",
-    title="Whisper Demo: Transcribe Audio",
-    description=(
-        f"Upload audio files to transcribe or translate them using the {MODEL_NAME} model."
-    ),
-    allow_flagging="never",
-    examples=[
-        ["./example.flac", "transcribe", False],
-        ["./example.flac", "transcribe", True],
-    ],
+    fn=transcribe_speech,
+    inputs=gr.Audio(sources="upload", type="filepath"),
+    outputs=gr.Textbox(),
 )
 
-# Create the Gradio tabbed interface for switching between modes
+# Creating the tabbed layout using Blocks
 demo = gr.Blocks()
 
 with demo:
     gr.TabbedInterface(
         [mic_transcribe, file_transcribe],
-        ["Transcribe Microphone", "Transcribe Audio File"]
+        ["Transcribe Microphone", "Transcribe Audio File"],
     )
 
-# Launch the app
+# Launch the app with debugging enabled
 if __name__ == "__main__":
-    demo.launch(debug=True, enable_queue=True, share=True)
+    demo.launch(debug=True, share=True)