debug

app.py

@@ -182,13 +182,15 @@ def http_bot(state):
         state.messages[-1][-1] = server_error_msg
         yield (state, state.to_gradio_chatbot_public()) + (enable_btn,) * 2
         return
+    tentative_answer = output
+    logging.info(f"Tentative Response: {tentative_answer}")
     state.messages[-1][-1] = state.messages[-1][-1][:-1]
     yield (state, state.to_gradio_chatbot_public()) + (disable_btn,) * 2
     
     # Step 2: Query-conditioned Caption
     state.append_message(state.roles[1], "# Query-conditioned Caption\n\n▌")
     try:
-        for generated_text in stream_response(mllm, cap_inputs, mllm_streamer, cao_prompt, mllm_sampling):
+        for generated_text in stream_response(mllm, cap_inputs, mllm_streamer, cap_prompt, mllm_sampling):
             output = generated_text[len(cap_prompt):].strip()
             state.messages[-1][-1] = "# Query-conditioned Caption\n\n" + output + "▌"
             yield (state, state.to_gradio_chatbot_public()) + (disable_btn,) * 2
@@ -198,6 +200,8 @@ def http_bot(state):
         state.messages[-1][-1] = server_error_msg
         yield (state, state.to_gradio_chatbot_public()) + (enable_btn,) * 2
         return
+    caption_text = output
+    logging.info(f"Query-conditioned Caption: {caption_text}")
     state.messages[-1][-1] = state.messages[-1][-1][:-1]
     yield (state, state.to_gradio_chatbot_public()) + (disable_btn,) * 2