Load and apply a trained Keras model¶
This example is from Stefan Wunsch (CERN IML TensoFlow and Keras workshop).
The code of this notebook shows how you can load and apply an already trained Keras model.
In [1]:
from os import environ
environ["KERAS_BACKEND"] = "tensorflow"
import numpy as np
import png
from keras.models import load_model
from os import listdir
import matplotlib.pyplot as plt
Load the model¶
Loading a Keras model needs only a single line of code, see below. After this call, the model is back in the same state you stored it at the training step either by the ModelCheckpoint or model.save(...).
In [2]:
model = load_model("mnist_keras_model.h5")
Apply the model¶
The application is done as shown in the testing phase of the training script. Simply call model.predict(inputs) on your data.
In [12]:
predictions = []
images = []
for f in sorted(listdir(".")):
if "mnist_example_" in f:
image = np.zeros((1, 28, 28, 1), dtype=np.uint8)
pngdata = png.Reader(open(f, 'rb')).asDirect()
for i_row, row in enumerate(pngdata[2]):
image[0, i_row, :, 0] = row
images.append(image)
prediction = np.argmax(model.predict(image))
predictions.append(prediction)
In [13]:
num_examples = len(images)
plt.figure(figsize=(num_examples*2, 2))
plt.rcParams.update({'axes.titlesize': 'xx-large'})
for i in range(num_examples):
plt.subplot(1, num_examples, i+1)
plt.axis('off')
plt.imshow(np.squeeze(images[i]), cmap="gray")
plt.title("{}".format(predictions[i]))
In [22]:
f = "mnist_my_digit_3.png"
image = np.zeros((1, 28, 28, 1), dtype=np.uint8)
pngdata = png.Reader(open(f, 'rb')).asDirect()
for i_row, row in enumerate(pngdata[2]):
image[0, i_row, :, 0] = row
prediction_vector = model.predict(image)
prediction = np.argmax(prediction_vector)
print (f"Model prediction for each class: {prediction_vector}")
print (f"Predicted digit: {prediction}")
plt.axis('off')
plt.imshow(np.squeeze(image), cmap="gray");
Model prediction for each class: [[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]] Predicted digit: 3
