File size: 3,081 Bytes
021e7cc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
import json

import gradio as gr
import spaces
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline

from format.format_output import format_output
from validate.validate_ingredients import validate_ingredients
from device.get_device_id import get_device_id

tokenizer = AutoTokenizer.from_pretrained("Ashikan/dut-recipe-generator")
model = AutoModelForCausalLM.from_pretrained("Ashikan/dut-recipe-generator")
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, device=get_device_id())


@spaces.GPU
def perform_model_inference(ingredients_list):
    for ingredient_index in range(len(ingredients_list)):
        ingredients_list[ingredient_index] = ingredients_list[ingredient_index].strip()

    input_text = '{"prompt": ' + json.dumps(ingredients_list)

    output = pipe(input_text, max_length=1024, temperature=0.1, do_sample=True, truncation=True)[0]["generated_text"]

    return format_output(output)


def generate_recipe(ingredients):
    ingredients_list = ingredients.lower().split(',')
    is_ingredients_valid = validate_ingredients(ingredients_list)

    if is_ingredients_valid:
        generated_text = perform_model_inference(ingredients_list)

        return {
            generated_recipe: gr.Markdown(value=generated_text, label="Generated Recipe",
                                          elem_id="recipe-container", visible=True)
        }
    else:
        error_text = "## Invalid ingredients. Please include at least 2 ingredients in a comma separated list. e.g. brown rice, onions, garlic"

        return {
            generated_recipe: gr.Markdown(value=error_text, elem_id="recipe-container", visible=True)
        }


with gr.Blocks(css="./css/styles.css") as recipegen:
    gr.Image("./assets/dut.png", interactive=False, show_share_button=False, show_download_button=False,
             show_fullscreen_button=False, show_label=False, elem_id="dut-logo", height=256)
    gr.Markdown("# Durban University Of Technology Recipe Generator", elem_id="header")
    gr.Markdown("### An AI Model Attempting To Produce Healthier, Diabetic-Friendly Recipes",
                elem_id="header-sub-heading")
    gr.Markdown("Start by entering a comma-separated list of ingredients below.", elem_id="header-instructions")
    with gr.Column() as column:
        user_ingredients = gr.Textbox(label="Ingredients", autofocus=True, max_lines=1, elem_id="ingredients-input")
        generate_button = gr.Button(value="Generate")
    with gr.Column():
        generated_recipe = gr.Markdown(visible=True)
    examples = gr.Examples(
        elem_id="examples",
        examples=[
            "sweet potato, mushrooms, cheese, garlic",
            "chicken breast, chili, onion, tomato, parmesan cheese",
            "strawberries, vanilla, honey, rolled oats, almonds, butter",
            "hake, spring onion, lemon"
        ],
        inputs=[user_ingredients]
    )

    generate_button.click(
        fn=generate_recipe,
        inputs=[user_ingredients],
        outputs=[generated_recipe]
    )

recipegen.launch(share=True)