Create app.py

app.py

@@ -0,0 +1,51 @@
+import streamlit as st
+from gradio_client import Client
+
+st.title("Whisper-JAX Speech-to-Text App")
+
+# Specify the API URL
+API_URL = "https://sanchit-gandhi-whisper-jax.hf.space"
+
+# Initialize the Gradio client with the API URL
+client = Client(API_URL)
+
+# Function to transcribe audio using the specified API endpoint
+def transcribe_audio(audio_path="temp.mp3", task="transcription", return_timestamps=False):
+    """Function to transcribe an audio file using the Whisper-JAX endpoint."""
+    # Making a synchronous call to the predict method
+    # Note that file needs to be passed as a tuple with the format: (filename, filedata)
+    text, runtime = client.predict(
+        ("file", open(audio_path, "rb")),  # Opening file in binary read mode
+        task,
+        return_timestamps,
+        api_name="/predict_1"  # Ensure this is the correct endpoint
+    )
+    return text, runtime
+
+# Streamlit widget to upload an audio file
+uploaded_file = st.file_uploader("Choose an audio file", type=['mp3', 'wav', 'ogg'])
+
+# Options for the task and timestamp inclusion
+task = st.radio("Choose a task", ["Transcription", "Translation"], index=0)
+return_timestamps = st.checkbox("Return timestamps with transcription")
+
+# Button to process the audio file
+if st.button("Transcribe Audio"):
+    if uploaded_file is not None:
+        # Save uploaded file temporarily
+        file_path = f"temp.mp3"
+        with open(file_path, "wb") as f:
+            f.write(uploaded_file.getbuffer())
+
+        # Call the transcribe function
+        try:
+            transcription = transcribe_audio()
+            st.write("Transcription:", transcription)
+        except Exception as e:
+            st.error(f"An error occurred during transcription: {str(e)}")
+        finally:
+            # Clean up the temporary file
+            import os
+            os.remove(file_path)
+    else:
+        st.error("Please upload an audio file to proceed.")