Fix: Load CPU-safe model for HF Space

app.py

@@ -6,6 +6,7 @@ from peft import PeftModel
 from pydub import AudioSegment
 import speech_recognition as sr
 import io
+from audiorecorder import audiorecorder  # Add at the top with other imports
 
 # Load model and tokenizer from local fine-tuned directory
 # Define base and adapter model paths
@@ -103,27 +104,53 @@ elif input_mode == "Text Phrase":
         except Exception as e:
             st.error(f"Error: {e}")
 
-elif input_mode == "Audio Upload":
-    uploaded_file = st.file_uploader("Upload audio file (WAV, MP3, M4A, MPEG)", type=["wav", "mp3", "m4a","mpeg"])
-    if uploaded_file:
-        st.audio(uploaded_file, format='audio/wav')
-        audio = AudioSegment.from_file(uploaded_file)
-        wav_io = io.BytesIO()
-        audio.export(wav_io, format="wav")
-        wav_io.seek(0)
-
-        recognizer = sr.Recognizer()
-        with sr.AudioFile(wav_io) as source:
-            audio_data = recognizer.record(source)
 
-        try:
-            text = recognizer.recognize_google(audio_data)
-            st.markdown(f"**Transcribed Text:** _{text}_")
-            values = extract_details_from_text(text)
-            if all(v is not None for v in values):
-                diagnosis = get_prediction(*values)
-                st.success(f"🩺 **Predicted Diagnosis:** {diagnosis}")
-            else:
-                st.warning("Could not extract complete information from audio.")
-        except Exception as e:
-            st.error(f"Audio processing error: {e}")
+elif input_mode == "Audio Upload":
+    audio_input_mode = st.radio("Choose audio input type:", ["Upload Audio File", "Record Audio"])
+
+    if audio_input_mode == "Upload Audio File":
+        uploaded_file = st.file_uploader("Upload audio file (WAV, MP3, M4A, MPEG)", type=["wav", "mp3", "m4a", "mpeg"])
+        if uploaded_file:
+            st.audio(uploaded_file, format='audio/wav')
+            audio = AudioSegment.from_file(uploaded_file)
+            wav_io = io.BytesIO()
+            audio.export(wav_io, format="wav")
+            wav_io.seek(0)
+
+            recognizer = sr.Recognizer()
+            with sr.AudioFile(wav_io) as source:
+                audio_data = recognizer.record(source)
+
+            try:
+                text = recognizer.recognize_google(audio_data)
+                st.markdown(f"**Transcribed Text:** _{text}_")
+                values = extract_details_from_text(text)
+                if all(v is not None for v in values):
+                    diagnosis = get_prediction(*values)
+                    st.success(f"🩺 **Predicted Diagnosis:** {diagnosis}")
+                else:
+                    st.warning("Could not extract complete information from audio.")
+            except Exception as e:
+                st.error(f"Audio processing error: {e}")
+
+    elif audio_input_mode == "Record Audio":
+        audio = audiorecorder("Click to record", "Recording...")
+        if len(audio) > 0:
+            st.audio(audio.tobytes(), format="audio/wav")
+            wav_io = io.BytesIO(audio.tobytes())
+
+            recognizer = sr.Recognizer()
+            with sr.AudioFile(wav_io) as source:
+                audio_data = recognizer.record(source)
+
+            try:
+                text = recognizer.recognize_google(audio_data)
+                st.markdown(f"**Transcribed Text:** _{text}_")
+                values = extract_details_from_text(text)
+                if all(v is not None for v in values):
+                    diagnosis = get_prediction(*values)
+                    st.success(f"🩺 **Predicted Diagnosis:** {diagnosis}")
+                else:
+                    st.warning("Could not extract complete information from recorded audio.")
+            except Exception as e:
+                st.error(f"Recording processing error: {e}")