Add audio transcription with Whisper

README.md

@@ -1,11 +1,17 @@
 ---
-title: Kab Hsf Test
-emoji: 👀
-colorFrom: yellow
-colorTo: red
-sdk: docker
-app_port: 7860
-pinned: false
-license: apache-2.0
-short_description: Testing the HSF environment
+title: Audio Transcriber
+emoji: 🎙️
+colorFrom: green
+colorTo: blue
+sdk: gradio
+app_file: app.py
+python_version: "3.10"
+short_description: Upload an audio file and get a transcription using Whisper.
+tags:
+  - audio
+  - speech-to-text
+  - whisper
+  - transcription
+models:
+  - openai/whisper-small
 ---

app.py

@@ -1,9 +1,43 @@
-from flask import Flask
-app = Flask(__name__)
+import gradio as gr
+import torch
+from transformers import pipeline
 
-@app.route("/")
-def hello():
-    return "<h1>Hello, World from Hugging Face Space!</h1>"
+# Load the Whisper model for speech recognition
+device = 0 if torch.cuda.is_available() else -1
+transcriber = pipeline(
+    "automatic-speech-recognition",
+    model="openai/whisper-small",
+    chunk_length_s=30,
+    device=device
+)
 
-if __name__ == "__main__":
-    app.run(host="0.0.0.0", port=7860)
+def transcribe_audio(audio_file):
+    """
+    Transcribe uploaded audio file.
+    Args:
+        audio_file (str): Path to the audio file
+    Returns:
+        str: Transcribed text
+    """
+    if audio_file is None:
+        return "Please upload an audio file."
+    
+    # Run transcription
+    result = transcriber(audio_file)
+    return result["text"]
+
+# Define the Gradio interface
+demo = gr.Interface(
+    fn=transcribe_audio,
+    inputs=gr.Audio(sources=["upload", "microphone"], type="filepath"),
+    outputs=gr.Textbox(label="Transcription", lines=8),
+    title="🎙️ Audio Transcription with Whisper",
+    description="Upload an audio file or record directly to transcribe it using OpenAI's Whisper model.",
+    examples=[
+        ["example.wav"]  # Optional: add a sample audio file named example.wav
+    ],
+    allow_flagging="never"
+)
+
+# Launch the app
+demo.launch()