-
Notifications
You must be signed in to change notification settings - Fork 300
Add EfficientNet Presets #1933
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.
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
Add EfficientNet Presets #1933
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @pkgoogle. Left a few NIT comments
expand_ratio=stackwise_expansion_ratios[i], | ||
se_ratio=squeeze_and_excite_ratio, | ||
activation=activation, | ||
project_activation=project_activation, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: project->projection
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
done
kernel_size=3, | ||
strides=1, | ||
activation="swish", | ||
project_activation=None, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
project->projection
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
done
depthwise_initializer=conv_kernel_initializer(), | ||
name=name + "dwconv", | ||
)(x) | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
remove blank line
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
done
@@ -0,0 +1,226 @@ | |||
import math |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
add instructions on how to run the checkpoint conversion script. see resnet script for example
"""
python ....
"""
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Did you mean this? The analogous convert_resnet.py does not have examples: convert_resnet.py
"metadata": { | ||
"description": ( | ||
"EfficientNet B0 model pre-trained on the ImageNet 1k dataset " | ||
"with RandAugment recipe." |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
what is the random augment recipie?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Here's the paper... should we link it?
Ports
efficientnet_b0_ra_imagenet
andefficientnet_b1_ft_imagenet
presets to EfficientNet and creates the scaffolding for further workTested Colab