create app.py

Using distil-whisper model to convert speech to text

app.py

@@ -0,0 +1,25 @@
+import gradio as gr
+from transformers import pipeline
+import spaces
+
+# Load the Whisper model from Hugging Face
+model = pipeline("automatic-speech-recognition", model="distil-whisper/distil-large-v3", chunk_length_s=30, device=0)
+
+# Function to process audio input and transcribe it
+@spaces.GPU
+def transcribe(audio):
+    # Load and preprocess the audio
+    transcription = model(audio,batch_size=1000, generate_kwargs={"task": "transcribe"}, return_timestamps=True)["text"]
+    return transcription
+
+
+# Gradio interface
+interface = gr.Interface(
+    fn=transcribe, 
+    inputs=gr.Audio(sources="microphone", type="filepath"),  
+    outputs="text",
+    title="Whisper Voice Transcription with Hugging Face"
+)
+
+# Launch the app
+interface.launch()