app.py

import gradio as gr
import os
import torch
import numpy

from timeit import default_timer as timer
from model import create_effnetb2_model
from typing import Tuple , Dict

# 1.Import and class names setup
class_names = ['apple_pie',
 'baby_back_ribs',
 'baklava',
 'beef_carpaccio',
 'beef_tartare',
 'beet_salad',
 'beignets',
 'bibimbap',
 'bread_pudding',
 'breakfast_burrito',
 'bruschetta',
 'caesar_salad',
 'cannoli',
 'caprese_salad',
 'carrot_cake',
 'ceviche',
 'cheese_plate',
 'cheesecake',
 'chicken_curry',
 'chicken_quesadilla',
 'chicken_wings',
 'chocolate_cake',
 'chocolate_mousse',
 'churros',
 'clam_chowder',
 'club_sandwich',
 'crab_cakes',
 'creme_brulee',
 'croque_madame',
 'cup_cakes',
 'deviled_eggs',
 'donuts',
 'dumplings',
 'edamame',
 'eggs_benedict',
 'escargots',
 'falafel',
 'filet_mignon',
 'fish_and_chips',
 'foie_gras',
 'french_fries',
 'french_onion_soup',
 'french_toast',
 'fried_calamari',
 'fried_rice',
 'frozen_yogurt',
 'garlic_bread',
 'gnocchi',
 'greek_salad',
 'grilled_cheese_sandwich',
 'grilled_salmon',
 'guacamole',
 'gyoza',
 'hamburger',
 'hot_and_sour_soup',
 'hot_dog',
 'huevos_rancheros',
 'hummus',
 'ice_cream',
 'lasagna',
 'lobster_bisque',
 'lobster_roll_sandwich',
 'macaroni_and_cheese',
 'macarons',
 'miso_soup',
 'mussels',
 'nachos',
 'omelette',
 'onion_rings',
 'oysters',
 'pad_thai',
 'paella',
 'pancakes',
 'panna_cotta',
 'peking_duck',
 'pho',
 'pizza',
 'pork_chop',
 'poutine',
 'prime_rib',
 'pulled_pork_sandwich',
 'ramen',
 'ravioli',
 'red_velvet_cake',
 'risotto',
 'samosa',
 'sashimi',
 'scallops',
 'seaweed_salad',
 'shrimp_and_grits',
 'spaghetti_bolognese',
 'spaghetti_carbonara',
 'spring_rolls',
 'steak',
 'strawberry_shortcake',
 'sushi',
 'tacos',
 'takoyaki',
 'tiramisu',
 'tuna_tartare',
 'waffles']


print(numpy.__version__)
from torchvision.models._api import WeightsEnum
from torch.hub import load_state_dict_from_url
def get_state_dict(self, *args, **kwargs):
    kwargs.pop("check_hash")
    return load_state_dict_from_url(self.url, *args, **kwargs)
WeightsEnum.get_state_dict = get_state_dict

# 2. Model annd transforms prepration
effnetb2_model  , effnet_b2_transforms = create_effnetb2_model(num_classes = 101, seed = 42)

# Load save weights
effnetb2_model.load_state_dict(
  torch.load(
    f='11_pretrained_effnet_feature_extractor_food101_fine_tune.pth',
    map_location = torch.device('cpu') # Load the model on CPU
  )
)
# 3.prediction function (predict())
def predict(img) -> Tuple[Dict,float] :

  start_time = timer()

  image =  effnet_b2_transforms(img).unsqueeze(0)


  effnetb2_model.eval()
  with torch.inference_mode():

    pred_probs = torch.softmax(effnetb2_model(image) , dim=1)
  pred_label_and_probs = {class_names[i]: float(pred_probs[0][i]) for i in range (len(class_names))}

  end_time = timer()
  pred_time = round(end_time - start_time , 4)

  return pred_label_and_probs , pred_time

### 4. Gradio app - our Gradio interface + launch command

title = 'FoodVision Big'
description = 'An FineTune last 4 Sequential layers of EfficientNetB2 model to classifiy 101 Food images '
article = 'created at PyTorch Model Deployment'

# Create example list

example_list = [['examples/'+ example] for example in os.listdir('examples')]
example_list

# create a gradio demo
demo = gr.Interface(fn=predict ,
                    inputs=gr.Image(type='pil'),
                    outputs=[gr.Label(num_top_classes = 3 , label= 'prediction'),
                            gr.Number(label= 'Prediction time (s)')],
                    examples = example_list,
                    title = title,
                    description = description,
                    article= article)

# Launch the demo
demo.launch(debug= False)