sdcubber

import pandas as pd
df_train = pd.read_csv('./KagglePlanetMCML.csv')
df_train.head()

sdcubber

import tensorflow as tf
IM_SIZE = 128

image_input = tf.keras.Input(shape=(IM_SIZE, IM_SIZE, 3), name='input_layer')

# Some convolutional layers
conv_1 = tf.keras.layers.Conv2D(32,
                                kernel_size=(3, 3),
                                padding='same',
                                activation='relu')(image_input)

sdcubber

print(model.summary())

sdcubber

model.compile(optimizer='adam',
              loss={'weather': 'categorical_crossentropy',
                    'ground': 'binary_crossentropy'})

sdcubber

import ast
import numpy as np
import math
import os
import random
from tensorflow.keras.preprocessing.image import img_to_array as img_to_array
from tensorflow.keras.preprocessing.image import load_img as load_img

def load_image(image_path, size):
    # data augmentation logic such as random rotations can be added here

sdcubber

callbacks = [
    tf.keras.callbacks.ModelCheckpoint('./model.h5', verbose=1)
]

model.fit_generator(generator=seq,
                    verbose=1, 
                    epochs=1,
                    use_multiprocessing=True,
                    workers=4,
                    callbacks=callbacks)

sdcubber

seq = KagglePlanetSequence('./KagglePlanetMCML.csv',
                       './data/train/',
                       im_size=IM_SIZE,
                       batch_size=32)

sdcubber

another_model = tf.keras.models.load_model('./model.h5')
another_model.fit_generator(generator=seq, verbose=1, epochs=1)

sdcubber

import tensorflow as tf
from tensorflow.python.saved_model import builder as saved_model_builder
from tensorflow.python.saved_model.signature_def_utils_impl import predict_signature_def
from tensorflow.python.saved_model import tag_constants
tf.keras.backend.set_learning_phase(0)    

# The export path contains the name and the version of the model
export_path = './PlanetModel/1'
model = tf.keras.models.load_model('./model.h5')
builder = saved_model_builder.SavedModelBuilder(export_path)

sdcubber

import requests
import json

image = img_to_array(load_img('./data/train/train_10001.jpg', target_size=(128,128))) / 255.
payload = {
  "instances": [{'input_image': image.tolist()}]
}
r = requests.post('http://localhost:9000/v1/models/PlanetModel:predict', json=payload)
json.loads(r.content)