visualkeras: autoencoder example#
An example showing the visualkeras function
used by a tf.keras.Model model.
9 # Authors: The scikit-plots developers
10 # SPDX-License-Identifier: BSD-3-Clause
pip install protobuf==5.29.4
14 import tensorflow as tf
15
16 # Clear any session to reset the state of TensorFlow/Keras
17 tf.keras.backend.clear_session()
encoder Model
21 encoder_input = tf.keras.Input(shape=(28, 28, 1), name="img")
22 x = tf.keras.layers.Conv2D(16, 3, activation="relu")(encoder_input)
23 x = tf.keras.layers.Conv2D(32, 3, activation="relu")(x)
24 x = tf.keras.layers.MaxPooling2D(3)(x)
25 x = tf.keras.layers.Conv2D(32, 3, activation="relu")(x)
26 x = tf.keras.layers.Conv2D(16, 3, activation="relu")(x)
27 encoder_output = tf.keras.layers.GlobalMaxPooling2D()(x)
28 encoder = tf.keras.Model(encoder_input, encoder_output, name="encoder")
29
30 # autoencoder Model
31 x = tf.keras.layers.Reshape((4, 4, 1))(encoder_output)
32 x = tf.keras.layers.Conv2DTranspose(16, 3, activation="relu")(x)
33 x = tf.keras.layers.Conv2DTranspose(32, 3, activation="relu")(x)
34 x = tf.keras.layers.UpSampling2D(3)(x)
35 x = tf.keras.layers.Conv2DTranspose(16, 3, activation="relu")(x)
36 decoder_output = tf.keras.layers.Conv2DTranspose(1, 3, activation="relu")(x)
37 autoencoder = tf.keras.Model(encoder_input, decoder_output, name="autoencoder")
38 autoencoder.summary()
Model: "autoencoder"
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type) ┃ Output Shape ┃ Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ img (InputLayer) │ (None, 28, 28, 1) │ 0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d (Conv2D) │ (None, 26, 26, 16) │ 160 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_1 (Conv2D) │ (None, 24, 24, 32) │ 4,640 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ max_pooling2d (MaxPooling2D) │ (None, 8, 8, 32) │ 0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_2 (Conv2D) │ (None, 6, 6, 32) │ 9,248 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_3 (Conv2D) │ (None, 4, 4, 16) │ 4,624 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ global_max_pooling2d │ (None, 16) │ 0 │
│ (GlobalMaxPooling2D) │ │ │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ reshape (Reshape) │ (None, 4, 4, 1) │ 0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_transpose │ (None, 6, 6, 16) │ 160 │
│ (Conv2DTranspose) │ │ │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_transpose_1 │ (None, 8, 8, 32) │ 4,640 │
│ (Conv2DTranspose) │ │ │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ up_sampling2d (UpSampling2D) │ (None, 24, 24, 32) │ 0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_transpose_2 │ (None, 26, 26, 16) │ 4,624 │
│ (Conv2DTranspose) │ │ │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ conv2d_transpose_3 │ (None, 28, 28, 1) │ 145 │
│ (Conv2DTranspose) │ │ │
└─────────────────────────────────┴────────────────────────┴───────────────┘
Total params: 28,241 (110.32 KB)
Trainable params: 28,241 (110.32 KB)
Non-trainable params: 0 (0.00 B)
Build the model with an explicit input shape
42 autoencoder.build(
43 input_shape=(None, 28, 28, 1)
44 ) # Batch size of None, shape (28, 28, 1)
45
46 # Create a dummy input tensor with a batch size of 1
47 dummy_input = tf.random.normal([1, 28, 28, 1]) # Batch size of 1, shape (28, 28, 1)
48 # Run the dummy input through the model to trigger shape calculation
49 encoder_output = autoencoder(dummy_input)
50 # Now check the output shape of the encoder
51 print("Output shape after running model with dummy input:", encoder_output.shape)
52
53 # Check each layer's output shape after building the model
54 for layer in encoder.layers:
55 if hasattr(layer, "output_shape"):
56 print(f"{layer.name} output shape: {layer.output_shape}")
57 if hasattr(layer, "output"):
58 print(f"{layer.name} shape: {layer.output.shape}")
Output shape after running model with dummy input: (1, 28, 28, 1)
img shape: (None, 28, 28, 1)
conv2d shape: (None, 26, 26, 16)
conv2d_1 shape: (None, 24, 24, 32)
max_pooling2d shape: (None, 8, 8, 32)
conv2d_2 shape: (None, 6, 6, 32)
conv2d_3 shape: (None, 4, 4, 16)
global_max_pooling2d shape: (None, 16)
61 from scikitplot import visualkeras
62
63 img_encoder = visualkeras.layered_view(
64 encoder,
65 text_callable="default",
66 # to_file="result_images/encoder.png",
67 save_fig=True,
68 save_fig_filename="encoder.png",
69 )
70 img_encoder
2025-06-27 09:09:35.492813: W scikitplot 140532842670976 utils_pil.py:203:load_font] Error loading system font: cannot open resource
2025-06-27 09:09:35.492891: W scikitplot 140532842670976 utils_pil.py:205:load_font] Falling back to PIL default font.
2025-06-27 09:09:35.493024: W scikitplot 140532842670976 layered.py:203:layered_view] The legend_text_spacing_offset parameter is deprecated andwill be removed in a future release.
<matplotlib.image.AxesImage object at 0x7fcf6c741c10>
73 img_autoencoder = visualkeras.layered_view(
74 autoencoder,
75 # to_file="result_images/autoencoder.png",
76 save_fig=True,
77 save_fig_filename="autoencoder.png",
78 )
79 img_autoencoder
2025-06-27 09:09:35.660838: W scikitplot 140532842670976 utils_pil.py:203:load_font] Error loading system font: cannot open resource
2025-06-27 09:09:35.660916: W scikitplot 140532842670976 utils_pil.py:205:load_font] Falling back to PIL default font.
2025-06-27 09:09:35.661049: W scikitplot 140532842670976 layered.py:203:layered_view] The legend_text_spacing_offset parameter is deprecated andwill be removed in a future release.
<matplotlib.image.AxesImage object at 0x7fcf5476ec90>
82 img_autoencoder_text = visualkeras.layered_view(
83 autoencoder,
84 min_z=1,
85 min_xy=1,
86 max_z=4096,
87 max_xy=4096,
88 scale_z=1,
89 scale_xy=1,
90 # font={"font_size": 14},
91 text_callable="default",
92 # to_file="result_images/autoencoder_text.png",
93 save_fig=True,
94 save_fig_filename="autoencoder_text.png",
95 overwrite=False,
96 add_timestamp=True,
97 verbose=True,
98 )
99 img_autoencoder_text
2025-06-27 09:09:35.798412: W scikitplot 140532842670976 utils_pil.py:203:load_font] Error loading system font: cannot open resource
2025-06-27 09:09:35.798497: W scikitplot 140532842670976 utils_pil.py:205:load_font] Falling back to PIL default font.
2025-06-27 09:09:35.798636: W scikitplot 140532842670976 layered.py:203:layered_view] The legend_text_spacing_offset parameter is deprecated andwill be removed in a future release.
[INFO] Saving path to: /home/circleci/repo/galleries/examples/visualkeras_CNN/result_images/autoencoder_text_20250627_090935Z.png
<matplotlib.image.AxesImage object at 0x7fcf543147d0>
Total running time of the script: (0 minutes 3.840 seconds)
Related examples
Visualkeras: Spam Classification Conv1D Dense Example
Visualkeras: Spam Classification Conv1D Dense Example
