import gradio as gr from transformers import AutoTokenizer, AutoModelForCausalLM import torch MODEL_ID = "distilgpt2" # fast + small + reliable on CPU tokenizer = AutoTokenizer.from_pretrained(MODEL_ID) model = AutoModelForCausalLM.from_pretrained(MODEL_ID) def respond(message, history): # Build a simple prompt from chat history prompt = "" for user, bot in history: prompt += f"User: {user}\nBot: {bot}\n" prompt += f"User: {message}\nBot:" inputs = tokenizer(prompt, return_tensors="pt") with torch.no_grad(): output_ids = model.generate( **inputs, max_new_tokens=120, do_sample=True, temperature=0.8, top_p=0.95, pad_token_id=tokenizer.eos_token_id, ) decoded = tokenizer.decode(output_ids[0], skip_special_tokens=True) reply = decoded.split("Bot:")[-1].strip() return reply demo = gr.ChatInterface( fn=respond, title="🤖 Fast Chat", description="Lightweight CPU-friendly chatbot running on Hugging Face Spaces." ) demo.launch()