app.py

import torch
from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM, AutoTokenizer

peft_model_id = f"Bsbell21/llm_instruction_generator"
model = AutoModelForCausalLM.from_pretrained(peft_model_id, return_dict=True, load_in_8bit=True, load_in_8bit_fp32_cpu_offload=True, device_map='auto')
tokenizer = AutoTokenizer.from_pretrained(peft_model_id)
# Load the Lora model
# model = PeftModel.from_pretrained(model, peft_model_id)

def input_from_text(text):
  return "<s>[INST]Use the provided input to create an instruction that could have been used to generate the response with an LLM.\n" + text + "[/INST]"

def get_instruction(text):
  inputs = mixtral_tokenizer(input_from_text(text), return_tensors="pt")

  outputs = merged_model.generate(
      **inputs,
      max_new_tokens=150,
      generation_kwargs={"repetition_penalty" : 1.7}
  )
  # print(mixtral_tokenizer.decode(outputs[0], skip_special_tokens=True))
  print(mixtral_tokenizer.decode(outputs[0], skip_special_tokens=True).split("[/INST]")[1])

if __name__ == "__main__":
    # make a gradio interface
    import gradio as gr

    gr.Interface(
        get_instruction,
        [
            gr.Textbox(lines=10, label="LLM Response"),
        ],
        gr.Textbox(label="LLM Predicted Prompt"),
        title="LLM-Prompt-Predictor",
        description="Prompt Predictor Based on LLM Response",
    ).launch()