text

app.py

@@ -226,26 +226,29 @@ def predict_category(que, input_image):
     return preds[0]
 
 
-def combine(audio, input_image):
-    que = transcribe_audio(audio)
-    # que = "What is the animal here?"
+def combine(audio, input_image, text_question=""):
+    if audio:
+        que = transcribe_audio(audio)
+    else:
+        que = text_question
 
     image = Image.fromarray(input_image).convert('RGB')
     category = predict_category(que, image)
-
     answer = predict_answer(0, que, image)
 
-    # print(category)
-
     tts = gTTS(answer)
     tts.save('answer.mp3')
+    
     return que, answer, 'answer.mp3'
 
-
-
-# Define the Gradio interface for recording audio and displaying the transcription
-model_interface = gr.Interface(fn=combine, inputs=[gr.Microphone(label="Ask your question"),gr.Image(label="Upload the image")], outputs=[gr.Text(label="Transcribed Question"), gr.Text(label="Answer"), gr.Audio(label="Audio Answer")])
-# image_upload_interface = gr.Interface(fn=upload_image, inputs=gr.Image(label="Upload the image"), outputs="text")
+# Define the Gradio interface for recording audio, text input, and image upload
+model_interface = gr.Interface(fn=combine, 
+                               inputs=[gr.Microphone(label="Ask your question"), 
+                                       gr.Image(label="Upload the image"),
+                                       gr.Textbox(label="Text Question")], 
+                               outputs=[gr.Text(label="Transcribed Question"), 
+                                        gr.Text(label="Answer"), 
+                                        gr.Audio(label="Audio Answer")])
 
 # Launch the Gradio interface
 model_interface.launch(debug=True)