Lower @spaces.GPU duration

app.py

@@ -196,7 +196,7 @@ def to_workflow(request: gr.Request, *args):
             yield value
 
 
-@spaces.GPU(duration=120)
+@spaces.GPU(duration=60)
 def run_workflow_local(*args):
     for value in run_workflow(*args):
         yield value
@@ -356,7 +356,7 @@ with gr.Blocks(
             status_text = f"""
             📍 Now in **local** mode, using ZeroGPU hardware<br>
             ⌛ Response time is around 2 minutes<br>
-            ✨ [Nomic](https://huggingface.co/nomic-ai/nomic-embed-text-v1.5) embeddings and [{model_id}](https://huggingface.co/{model_id}) LLM<br>
+            ✨ [Nomic](https://huggingface.co/nomic-ai/nomic-embed-text-v1.5) embeddings and [{model_id}](https://huggingface.co/{model_id})<br>
             🏠 See the project's [GitHub repository](https://github.com/jedick/R-help-chat)
             """
         return status_text
@@ -404,7 +404,7 @@ with gr.Blocks(
                 example_questions = [
                     # "What is today's date?",
                     "Summarize emails from the last two months",
-                    "What plotmath examples have been discussed?",
+                    "How to use plotmath?",
                     "When was has.HLC mentioned?",
                     "Who reported installation problems in 2023-2024?",
                 ]

graph.py

@@ -223,9 +223,9 @@ def BuildGraph(
         if is_local:
             # Don't include the system message here because it's defined in ToolCallingLLM
             messages = state["messages"]
-            print_message_summaries(messages, "--- query: before normalization ---")
+            # print_message_summaries(messages, "--- query: before normalization ---")
             messages = normalize_messages(messages)
-            print_message_summaries(messages, "--- query: after normalization ---")
+            # print_message_summaries(messages, "--- query: after normalization ---")
         else:
             messages = [SystemMessage(query_prompt(compute_mode))] + state["messages"]
         response = query_model.invoke(messages)
@@ -236,13 +236,13 @@ def BuildGraph(
         """Generates an answer with the chat model"""
         if is_local:
             messages = state["messages"]
-            print_message_summaries(messages, "--- generate: before normalization ---")
+            # print_message_summaries(messages, "--- generate: before normalization ---")
             messages = normalize_messages(messages)
             # Add the system message here because we're not using tools
             messages = [
                 SystemMessage(generate_prompt(with_tools=False, think=False))
             ] + messages
-            print_message_summaries(messages, "--- generate: after normalization ---")
+            # print_message_summaries(messages, "--- generate: after normalization ---")
         else:
             messages = [SystemMessage(generate_prompt())] + state["messages"]
         response = generate_model.invoke(messages)