Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| from smolagents import CodeAgent,HfApiModel # adjust the import to your actual smolagents module | |
| # Step 1: Set up your smolagents agent | |
| def create_agent(): | |
| """ | |
| Initialize and return the agent. | |
| Adjust parameters like model type or configuration as needed. | |
| """ | |
| # For example, we initialize an Agent with a sample model | |
| agent = CodeAgent(tools=[], model=HfApiModel(model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud/')) # Change arguments as per your agent configuration | |
| return agent | |
| # Create the agent instance once so that it persists across user interactions | |
| agent = create_agent() | |
| def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p): | |
| """ | |
| This function builds the conversation history, calls the smolagents agent, | |
| and streams the response back to Gradio. | |
| """ | |
| # Build the conversation messages list, starting with the system prompt. | |
| messages = [{"role": "system", "content": system_message}] | |
| for user_msg, assistant_msg in history: | |
| if user_msg: | |
| messages.append({"role": "user", "content": user_msg}) | |
| if assistant_msg: | |
| messages.append({"role": "assistant", "content": assistant_msg}) | |
| # Add the latest user input. | |
| messages.append({"role": "user", "content": message}) | |
| # Initialize an empty response. | |
| complete_response = "" | |
| # Step 2: Call the agent's chat_completion method. | |
| # If your smolagents agent supports streaming (i.e. yielding tokens as they are generated), | |
| # this loop will yield partial responses to update the UI. | |
| # If streaming is not supported, you can simply do: | |
| # complete_response = agent.chat_completion(messages, max_tokens, temperature, top_p) | |
| # yield complete_response | |
| # for token in agent.chat_completion( | |
| # messages, | |
| # max_tokens=max_tokens, | |
| # temperature=temperature, | |
| # top_p=top_p, | |
| # stream=True # set to False if your agent does not support streaming | |
| # ): | |
| # complete_response += token | |
| complete_response=agent.run(messages) | |
| yield complete_response | |
| # Step 3: Create the Gradio ChatInterface. | |
| demo = gr.ChatInterface( | |
| fn=respond, | |
| additional_inputs=[ | |
| gr.Textbox(value="You are a friendly Chatbot.", label="System message"), | |
| gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"), | |
| gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"), | |
| gr.Slider( | |
| minimum=0.1, | |
| maximum=1.0, | |
| value=0.95, | |
| step=0.05, | |
| label="Top-p (nucleus sampling)" | |
| ), | |
| ], | |
| ) | |
| # Step 4: Launch the Gradio app. | |
| if __name__ == "__main__": | |
| demo.launch() | |