app.py

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

peft_model_id = f"jmartin233/bloom-1b7-lora-reading-comprehension"
config = PeftConfig.from_pretrained(peft_model_id)
model = AutoModelForCausalLM.from_pretrained(
    config.base_model_name_or_path,
    return_dict=True,
    load_in_8bit=False,
    device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)


# Load the Lora model
model = PeftModel.from_pretrained(model, peft_model_id)

def make_inference(person, location, grammar, level):
  
  batch = tokenizer(f"""
    Below is a set of requirements for a short passage of English. Please write a passage that meets these requirements:

    ### Requirements:
    person: {person}
    location: {location}. 
    grammar: {grammar}
    level: {level} 

    ### Passage:
    Passage:""",
    return_tensors='pt')

  with torch.cuda.amp.autocast():
    output_tokens = model.generate(**batch, max_new_tokens=100)

  return tokenizer.decode(output_tokens[0], skip_special_tokens=True)



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

    gr.Interface(
        make_inference,
        [
            gr.inputs.Textbox(lines=2, label="Someone's name"),
            gr.inputs.Textbox(lines=2, label="A location they might visit"),
            gr.inputs.Textbox(lines=2, label="A type of grammar to use"),
            gr.inputs.Textbox(lines=2, label="The level of English to use (beginner, intermediate, advanced))"),

        ],
        gr.outputs.Textbox(label="Passage"),
        title="Reading Comprehension",
        description="A generative model that generates simple texts for testing reading comprehension.",
    ).launch()