app.py

import gradio as gr

from transformers import TextIteratorStreamer
from threading import Thread
import spaces

@spaces.GPU
def food_safety_query(prompt):
    from unsloth import FastLanguageModel
    # Model configuration
    model_name = "yasserrmd/food_safety"  
    max_seq_length = 512
    load_in_4bit = True  
    dtype = "fp16"  
    
    # Load model and tokenizer
    model, tokenizer = FastLanguageModel.from_pretrained(
        model_name=model_name,
        max_seq_length=max_seq_length,
        dtype=dtype,
        load_in_4bit=load_in_4bit,
    )
    FastLanguageModel.for_inference(model)  # Enable faster inference




    messages = [
        {"role": "system", "content": "You are an AI assistant with expertise in food safety. Your primary goal is to provide precise, actionable, and scientifically accurate responses to queries about food safety practices, standards, and regulations. Focus on offering guidance based on global food safety standards, such as HACCP, ISO 22000, and FDA guidelines. Your responses should address hygiene, contamination prevention, food handling, storage, production processes, and safety protocols with practical and specific advice."},
        {"role": "user", "content": prompt},
    ]

    # Apply chat template and obtain inputs
    inputs = tokenizer.apply_chat_template(
        messages,
        tokenize=True,
        add_generation_prompt=True,
        return_tensors="pt",
    ).to("cuda")

    # Check if inputs is a dictionary and contains 'input_ids'
    if isinstance(inputs, dict) and "input_ids" in inputs:
        input_ids = inputs["input_ids"]
    else:
        input_ids = inputs  # Assume inputs is already a tensor

    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)

    generation_kwargs = dict(
        input_ids=input_ids,
        max_new_tokens=2048,
        use_cache=True,
        temperature=1.5,
        min_p=0.1,
        streamer=streamer
    )

    thread = Thread(target=model.generate, kwargs=generation_kwargs)
    thread.start()
    output = ""
    for new_text in streamer:
        output += new_text
        yield output



with gr.Blocks() as app:
    gr.Markdown("""# Food Safety App\nEnter your questions related to food safety, and the assistant will provide detailed responses.
    """)
    prompt = gr.Textbox(label="Enter your query:", placeholder="E.g., What hygiene practices should milk-producing factories follow?")
    submit_button = gr.Button("Get Response")
    response = gr.Markdown()  # Changed response to Markdown for better display

    submit_button.click(
        fn=food_safety_query,
        inputs=[prompt],
        outputs=[response]
    )

# Launch the app
if __name__ == "__main__":
    app.launch(debug=True)