Update app.py

app.py

@@ -43,24 +43,24 @@ The model was trained on Thai audio recordings with the following sentences: \n
 
 
 
-def predict(file_path):
+def predict(file_upload,microphone):
     max_length = 100000
-    # file_path = filepath(file_upload)
-    # if (microphone is not None) and (file_upload is not None):
-    #     warn_output = (
-    #         "WARNING: You've uploaded an audio file and used the microphone. "
-    #         "The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
-    #     )
-
-    # elif (microphone is None) and (file_upload is None):
-    #     return "ERROR: You have to either use the microphone or upload an audio file"
-    # if(microphone is not None):
-    #     file_path = filepath(microphone)
-    # if(file_upload is not None):
-    #     file_path = filepath(microphone)
+    file_path =file_upload
+    if (microphone is not None) and (file_upload is not None):
+        warn_output = (
+            "WARNING: You've uploaded an audio file and used the microphone. "
+            "The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
+        )
+
+    elif (microphone is None) and (file_upload is None):
+        return "ERROR: You have to either use the microphone or upload an audio file"
+    if(microphone is not None):
+        file_path = microphone
+    if(file_upload is not None):
+        file_path = file_upload
     model.eval()
     with torch.no_grad():
-        wav_data, _ = sf.read(file_path.name)
+        wav_data, _ = sf.read(file_path)
         inputs = processor(wav_data, sampling_rate=16000, return_tensors="pt", padding=True)
 
         input_values = inputs.input_values.squeeze(0)  
@@ -75,10 +75,13 @@ def predict(file_path):
         logits = logits.squeeze()
         predicted_class_id = torch.argmax(logits, dim=-1).item()
 
-    return predicted_class_id
+    return "You probably have SP" if predicted_class_id == 1 else "You probably don't have SP"
 gr.Interface(
     fn=predict,
-    inputs=gr.inputs.Audio(type="file", label="Upload Audio File"),
+    inputs=[
+        gr.inputs.Audio(source="upload", type="filepath", optional=True),
+        gr.inputs.Audio(source="microphone", type="filepath", optional=True),
+    ],
     outputs="text",
     title=title,
     description=description,