Spaces:

fffiloni
/

llm-from-image

Sleeping

fffiloni commited on Feb 4

Commit

cc4f886

•

1 Parent(s): 425d180

Update app.py

Files changed (1) hide show

app.py CHANGED Viewed

@@ -78,7 +78,7 @@ def get_llm_idea(user_prompt):
     agent_maker_sys = f"""
 You are an AI whose job is to help users create their own chatbot whose personality will reflect the character or scene from an image described by users.
 In particular, you need to respond succintly in a friendly tone, write a system prompt for an LLM, a catchy title for the chatbot, and a very short example user input. Make sure each part is included.
-The system prompt will not mention any image provided. But You can include depicted details about the character if it makes sense for a more profound personality.
 For example, if a user says, "a picture of a man in a black suit and tie riding a black dragon", first do a friendly response, then add the title, system prompt, and example user input.
 Immediately STOP after the example input. It should be EXACTLY in this format:

     agent_maker_sys = f"""
 You are an AI whose job is to help users create their own chatbot whose personality will reflect the character or scene from an image described by users.
 In particular, you need to respond succintly in a friendly tone, write a system prompt for an LLM, a catchy title for the chatbot, and a very short example user input. Make sure each part is included.
+The system prompt will not mention any image provided. But You can include provided additional details about the character to the System Prompt, if it makes sense for a more sophisticated LLM persona.
 For example, if a user says, "a picture of a man in a black suit and tie riding a black dragon", first do a friendly response, then add the title, system prompt, and example user input.
 Immediately STOP after the example input. It should be EXACTLY in this format: