In [ ]:

%matplotlib inline

%matplotlib inline

Tensorflow image classification using pre-trained model example I¶

An example of a tensorflow pre-trained network that classifies indoor scenes images, where all layers are trained. For smaller datasets or datasets similar to the dataset the base model was trained on, it is advisable to freeze pre-trained network, and only train custom layers.

In [ ]:

import openml
import openml_tensorflow
import pandas as pd
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import tensorflow as tf
from tensorflow.keras.applications import EfficientNetB0
from tensorflow.keras import optimizers, Model
from tensorflow.keras.layers import Dense, GlobalAveragePooling2D

import openml import openml_tensorflow import pandas as pd from tensorflow.keras.preprocessing.image import ImageDataGenerator import tensorflow as tf from tensorflow.keras.applications import EfficientNetB0 from tensorflow.keras import optimizers, Model from tensorflow.keras.layers import Dense, GlobalAveragePooling2D

In [ ]:

openml.config.apikey = 'KEY' # Paste your API key here

openml.config.apikey = 'KEY' # Paste your API key here

In [ ]:

openml_tensorflow.config.epoch = 1 #  small epoch for test runs

datagen = ImageDataGenerator()
openml_tensorflow.config.datagen = datagen
openml_tensorflow.config.dir = openml.config.get_cache_directory()+'/datasets/45923/Images/'
openml_tensorflow.config.x_col = "Filename"
openml_tensorflow.config.y_col = 'Class_encoded'
openml_tensorflow.config.datagen = datagen
openml_tensorflow.config.batch_size = 2
openml_tensorflow.config.class_mode = "categorical"
openml_tensorflow.config.perform_validation = True

kwargs = {
    'callbacks': tf.keras.callbacks.EarlyStopping(monitor='loss', patience=0),
    'verbose': 2
}
openml_tensorflow.config.kwargs = kwargs

IMG_SIZE = 128
NUM_CLASSES = 67
base_learning_rate = 0.0001

# Example pre-trained model   
base_model = EfficientNetB0(input_shape=(IMG_SIZE, IMG_SIZE, 3),
                        weights="imagenet",
                        include_top=False)
x = base_model.output
x = GlobalAveragePooling2D()(x)
predictions = Dense(NUM_CLASSES, activation='softmax')(x)
model = Model(inputs=base_model.input, outputs=predictions)
model.compile(optimizers.Adam(learning_rate=4e-4),
                loss='categorical_crossentropy',
                metrics=['AUC'])

openml_tensorflow.config.epoch = 1 # small epoch for test runs datagen = ImageDataGenerator() openml_tensorflow.config.datagen = datagen openml_tensorflow.config.dir = openml.config.get_cache_directory()+'/datasets/45923/Images/' openml_tensorflow.config.x_col = "Filename" openml_tensorflow.config.y_col = 'Class_encoded' openml_tensorflow.config.datagen = datagen openml_tensorflow.config.batch_size = 2 openml_tensorflow.config.class_mode = "categorical" openml_tensorflow.config.perform_validation = True kwargs = { 'callbacks': tf.keras.callbacks.EarlyStopping(monitor='loss', patience=0), 'verbose': 2 } openml_tensorflow.config.kwargs = kwargs IMG_SIZE = 128 NUM_CLASSES = 67 base_learning_rate = 0.0001 # Example pre-trained model base_model = EfficientNetB0(input_shape=(IMG_SIZE, IMG_SIZE, 3), weights="imagenet", include_top=False) x = base_model.output x = GlobalAveragePooling2D()(x) predictions = Dense(NUM_CLASSES, activation='softmax')(x) model = Model(inputs=base_model.input, outputs=predictions) model.compile(optimizers.Adam(learning_rate=4e-4), loss='categorical_crossentropy', metrics=['AUC'])

In [ ]:

# Download the OpenML task for the Indoor Scenes dataset.
task = openml.tasks.get_task(362070)#   3 fold cross validation

model = model

# Run the Keras model on the task (requires an API key).
run = openml.runs.run_model_on_task(model, task, avoid_duplicate_runs=False)

# Download the OpenML task for the Indoor Scenes dataset. task = openml.tasks.get_task(362070)# 3 fold cross validation model = model # Run the Keras model on the task (requires an API key). run = openml.runs.run_model_on_task(model, task, avoid_duplicate_runs=False)

Note: If you want to publish the run with the onnx file, then you must call openml_tensorflow.add_onnx_to_run() immediately before run.publish(). When you publish, onnx file of last trained model is uploaded. Careful to not call this function when another run_model_on_task is called in between, as during publish later, only the last trained model (from last run_model_on_task call) is uploaded.

In [ ]:

run = openml_tensorflow.add_onnx_to_run(run)

run.publish()

print('URL for run: %s/run/%d?api_key=%s' % (openml.config.server, run.run_id, openml.config.apikey))

run = openml_tensorflow.add_onnx_to_run(run) run.publish() print('URL for run: %s/run/%d?api_key=%s' % (openml.config.server, run.run_id, openml.config.apikey))

Optional: Visualize model in netron

In [ ]:

from urllib.request import urlretrieve

published_run = openml.runs.get_run(run.run_id)
url = 'https://api.openml.org/data/download/{}/model.onnx'.format(published_run.output_files['onnx_model'])

file_path, _ = urlretrieve(url, 'model.onnx')

import netron
# Visualize the ONNX model using Netron
netron.start(file_path)

from urllib.request import urlretrieve published_run = openml.runs.get_run(run.run_id) url = 'https://api.openml.org/data/download/{}/model.onnx'.format(published_run.output_files['onnx_model']) file_path, _ = urlretrieve(url, 'model.onnx') import netron # Visualize the ONNX model using Netron netron.start(file_path)