vgg-48-get-data.py

import t4
import pandas as pd


# download the data
img_dir = 'images_cropped/'
metadata_filepath = 'X_meta.csv'
open_fruits = t4.Package.browse('quilt/open_fruit', 's3://quilt-example')
open_fruits['training_data/X_meta.csv'].fetch(metadata_filepath)
open_fruits['images_cropped'].fetch(img_dir)


# get X and y values for flow_from_directory
X_meta = pd.read_csv(metadata_filepath)
X = X_meta[['CroppedImageURL']].values
y = X_meta['LabelName'].values


# define data generators
train_datagen = ImageDataGenerator(
    rotation_range=40,
    width_shift_range=0.2,
    height_shift_range=0.2,
    rescale=1/255,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,
    fill_mode='nearest',
    validation_split=0.2
)
test_datagen = ImageDataGenerator(
    rescale=1/255,
)
train_generator = train_datagen.flow_from_directory(
    img_dir,
    target_size=(48, 48),
    batch_size=batch_size,
    class_mode='categorical',
    subset='training',
)
validation_generator = train_datagen.flow_from_directory(
    img_dir,
    target_size=(48, 48),
    batch_size=batch_size,
    class_mode='categorical',
    subset='validation'
)