app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "brdemorin/Phi3_80_steps_v2"

# Update the following parameters as necessary for your GPTQ model
gptq_parameters = {
    "load_in_8bit": False,
    "device_map": "auto",
    "trust_remote_code": True,  # This is often needed for custom models
}

# Load the model with GPTQ parameters
model = AutoModelForCausalLM.from_pretrained(model_name, **gptq_parameters)
tokenizer = AutoTokenizer.from_pretrained(model_name)

def chat(input_text):
    inputs = tokenizer(input_text, return_tensors="pt")
    outputs = model.generate(**inputs, max_new_tokens=50)
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response

iface = gr.Interface(fn=chat, inputs="text", outputs="text", title="Chat with Phi3_80")

iface.launch()