Skip to content

Corrections for gradient centralization example. #2113

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Corrections for gradient centralization example. #2113

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
48 changes: 26 additions & 22 deletions examples/vision/gradient_centralization.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,10 +2,11 @@
Title: Gradient Centralization for Better Training Performance
Author: [Rishit Dagli](https://github.com/Rishit-dagli)
Date created: 06/18/21
Last modified: 07/25/23
Last modified: 05/29/25
Description: Implement Gradient Centralization to improve training performance of DNNs.
Accelerator: GPU
Converted to Keras 3 by: [Muhammad Anas Raza](https://anasrz.com)
Debugged by: [Alberto M. Esmorís](https://github.com/albertoesmp)
"""

"""
Expand Down Expand Up @@ -122,27 +123,28 @@ def prepare(ds, shuffle=False, augment=False):
In this section we will define a Convolutional neural network.
"""

model = keras.Sequential(
[
layers.Input(shape=input_shape),
layers.Conv2D(16, (3, 3), activation="relu"),
layers.MaxPooling2D(2, 2),
layers.Conv2D(32, (3, 3), activation="relu"),
layers.Dropout(0.5),
layers.MaxPooling2D(2, 2),
layers.Conv2D(64, (3, 3), activation="relu"),
layers.Dropout(0.5),
layers.MaxPooling2D(2, 2),
layers.Conv2D(64, (3, 3), activation="relu"),
layers.MaxPooling2D(2, 2),
layers.Conv2D(64, (3, 3), activation="relu"),
layers.MaxPooling2D(2, 2),
layers.Flatten(),
layers.Dropout(0.5),
layers.Dense(512, activation="relu"),
layers.Dense(1, activation="sigmoid"),
]
)
def make_model():
return keras.Sequential(
[
layers.Input(shape=input_shape),
layers.Conv2D(16, (3, 3), activation="relu"),
layers.MaxPooling2D(2, 2),
layers.Conv2D(32, (3, 3), activation="relu"),
layers.Dropout(0.5),
layers.MaxPooling2D(2, 2),
layers.Conv2D(64, (3, 3), activation="relu"),
layers.Dropout(0.5),
layers.MaxPooling2D(2, 2),
layers.Conv2D(64, (3, 3), activation="relu"),
layers.MaxPooling2D(2, 2),
layers.Conv2D(64, (3, 3), activation="relu"),
layers.MaxPooling2D(2, 2),
layers.Flatten(),
layers.Dropout(0.5),
layers.Dense(512, activation="relu"),
layers.Dense(1, activation="sigmoid"),
]
)

"""
## Implement Gradient Centralization
Expand Down Expand Up @@ -216,6 +218,7 @@ def on_epoch_end(self, batch, logs={}):
"""

time_callback_no_gc = TimeHistory()
model = make_model()
model.compile(
loss="binary_crossentropy",
optimizer=RMSprop(learning_rate=1e-4),
Expand All @@ -241,6 +244,7 @@ def on_epoch_end(self, batch, logs={}):
"""

time_callback_gc = TimeHistory()
model = make_model()
model.compile(loss="binary_crossentropy", optimizer=optimizer, metrics=["accuracy"])

model.summary()
Expand Down
Loading