app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel, PeftConfig

# Load the PEFT configuration, base model, and tokenizer
config = PeftConfig.from_pretrained("SahilCarterr/Llama-2-7B-Chat-PEFT")
base_model = AutoModelForCausalLM.from_pretrained("TheBloke/Llama-2-7b-Chat-GPTQ", device_map='auto')
model = PeftModel.from_pretrained(base_model, "SahilCarterr/Llama-2-7B-Chat-PEFT")
tokenizer = AutoTokenizer.from_pretrained("SahilCarterr/Llama-2-7B-Chat-PEFT")

def respond(message, history, system_message, max_tokens, temperature, top_p):
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    messages.append({"role": "user", "content": message})
    
    # Encode the input
    inputs = tokenizer(message, return_tensors="pt").input_ids.to('cuda')
    # Generate the response using the model
    outputs = model.generate(inputs, max_new_tokens=max_tokens, do_sample=True, temperature=temperature, top_p=top_p)
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)

    yield response

demo = gr.ChatInterface(
    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)",
        ),
    ],
)

if __name__ == "__main__":
    demo.launch()