AdamP-Tensorflow

AdamP-Tensorflow copied to clipboard

AdamP Optimizer — Unofficial TensorFlow Implementation

"Slowing Down the Weight Norm Increase in Momentum-based Optimizers"

Implemented by Junho Kim

[Paper] [Project page] [Official Pytorch]

Validation

I have checked that the code is working, but I couldn't confirm if the performance is the same as the offical code.

Usage

Usage is exactly same as tf.keras.optimizers library!

from adamp_tf import AdamP
from sgdp_tf import SGDP

optimizer_adamp = AdamP(learning_rate=0.001, beta_1=0.9, beta_2=0.999, weight_decay=1e-2)
optimizer_sgdp = SGDP(learning_rate=0.1, weight_decay=1e-5, momentum=0.9, nesterov=True)

• Do not use with tf.nn.scale_regularization_loss. Use the weight_decay argument.

Arguments

SGDP and AdamP share arguments with tf.keras.optimizers.SGD and tf.keras.optimizers.Adam. There are two additional hyperparameters; we recommend using the default values.

• delta : threhold that determines whether a set of parameters is scale invariant or not (default: 0.1)
• wd_ratio : relative weight decay applied on scale-invariant parameters compared to that applied on scale-variant parameters (default: 0.1)

Both SGDP and AdamP support Nesterov momentum.

• nesterov : enables Nesterov momentum (default: False)

How to cite

@article{heo2020adamp,
    title={Slowing Down the Weight Norm Increase in Momentum-based Optimizers},
    author={Heo, Byeongho and Chun, Sanghyuk and Oh, Seong Joon and Han, Dongyoon and Yun, Sangdoo and Uh, Youngjung and Ha, Jung-Woo},
    year={2020},
    journal={arXiv preprint arXiv:2006.08217},
}