workingcode

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_name = "microsoft/DialoGPT-large"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

history_file = "chat_history.pth"

try:
    chat_history_ids = torch.load(history_file).to(device)
except (FileNotFoundError, RuntimeError):
    chat_history_ids = None


def generate_response(user_input, temperature=0.6, top_k=50, top_p=0.9, max_history=1024):
    global chat_history_ids

    input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors="pt").to(device)

    if chat_history_ids is not None:
        input_ids = torch.cat([chat_history_ids, input_ids], dim=-1)

    # Trim history to avoid exceeding model limits
    if input_ids.shape[-1] > max_history:
        input_ids = input_ids[:, -max_history:]

    attention_mask = torch.ones_like(input_ids, dtype=torch.long).to(device)

    response_ids = model.generate(
        input_ids,
        attention_mask=attention_mask,
        max_length=min(input_ids.shape[-1] + 50, 2048),  # Reduced max response length
        pad_token_id=tokenizer.eos_token_id,
        temperature=temperature,
        top_k=top_k,
        top_p=top_p,
        do_sample=True,
        no_repeat_ngram_size=2,
    )

    response_text = tokenizer.decode(response_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True)

    chat_history_ids = response_ids[:, -max_history:]
    torch.save(chat_history_ids, history_file)

    return response_text


if __name__ == "__main__":
    print("Model loaded successfully")
    print("Chatbot is running! Type 'exit' to quit.")

    while True:
        user_input = input("You: ")
        if user_input.lower() == "exit":
            print("Chatbot: Goodbye!")
            break

        response = generate_response(user_input)
        print("Chatbot:", response)