Skip to content

BariscanBozkurt/Gradient-Origin-Networks-Julia

Repository files navigation

Gradient-Origin-Networks-Julia

Reimplementation of Gradient Origin Networks [1] in Julia with Knet Framework (Koç University Deep Learning Framework).

This repo contains the Julia implementation of the paper Gradient Origin Networks [1] with the Knet Framework.

Implementation Environment Details

-> Julia Version Info: Version 1.6.3 (2021-09-23)

-> Platform Info : "OS: Linux (x86_64-pc-linux-gnu) CPU: Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz, GPU : CuDevice(0): Tesla T4"

-> Required Julia Packages : Specified in requirements.txt file

Implementation Notes

-> src folder includes Julia implementations for GON, Variational GON and Implicit GON models as well as some other model implementations (AE, VAE, etc.) to compare.

-> Pytorch Debug Files : The codes in Pytorch_Debug_Files are based on the published GON implementation [2] which provides loggers for train and test set loss values for each epoch. I use them to compare my Julia implementation with Pytorch implementation.

-> Experiment Notebooks : "Experimental_Notebooks" and "Experimental_Notebooks_Deprecated" files includes my debugging and experimenting notebooks. However, these notebooks will be deleted soon since I include the finalized version of the codes in "src" folder.

TODO

-> Optimization performance check for Variational GON.

Datasets

We performed experiments on MNIST, Fashion-MNIST, CIFAR-10 and subsampled COIL-20 image databases. The results are given in the following sections.

Results

We trained GON and Variational GON models with 4 Transposed Convolution Layers having 4x4 filters followed by Batch Normalization Layer and ELU activation functions. We trained Implicit GON with 5 linear layers of mapping sizes [34->256->256->256->256->1] followed by a sinusoidal activation function except the last layer. We used SIREN layer weight initialization [3].

GON

Model\Dataset MNIST Fashion-MNIST CIFAR-10 COIL-20
GON (ngf:16,nz:256) 0.1176/.01423 0.6043/0.6427 7.6046/7.7152 1.3784/2.5973
GON Detached (ngf:16, nz:256) 2.102/2.463 3.996/4.123 8.685/8.728 4.428/5.410
GON (ngf:32, nz:256) 0.0642/0.0884 0.4482/0.5184 6.871/7.260 0.5529/3.146
GON (ngf:16, nz:512) 0.542/0.0903 0.1233/0.132 3.346/3.413 -
Conv AE (ngf:16, nz:256) 0.0815/0.1352 0.5223/0.5755 7.382/7.542 0.942/4.537
Conv AE (ngf:32, nz:256) 0.0438/0.0775 0.3899/0.4804 6.496/6.959 0.4344/4.0801

Table 1: Summed Squared Error Losses on Train and Test Sets After 500 Epochs of Training

Implicit GON

Model\Dataset MNIST Fashion-MNIST COIL-20
Implicit - GON 0.00530/0.00535 0.0087/0.0089 0.0072/0.115

Table 2: Mean Squared Error Losses on Train and Test Sets (MNIST and Fashion-MNIST trained for 100 Epochs whereas COIL-20 trained for 500 Epochs)

Variational GON

Model\Dataset MNIST Fashion-MNIST CIFAR-10 COIL-20
V-GON (ngf:16, nz:256) 8.5681/8.9609 9.9733/10.178 43.269/43.812 8.917/11.427
VAE (ngf:8, nz:256) 6.859/7.3373 9.043/9.205 36.827/37.018 10.386/11.855
VAE (ngf:16, nz:256) 6.057/6.554 8.178/8.429 35.176/35.622 8.1/9.31

Table 3: Summed Squared Error Losses on Train and Test Sets After 1000 Epochs of Training

Visaual Results

GON

GON (ngf:16, nz:256) Test Set Reconstruction Examples

MNIST

Sample From Test Set

Fashion MNIST

Sample From Test Set

CIFAR-10

Sample From Test Set

COIL-20

Sample From Test Set

GON (ngf:16, nz:512) Test Set Reconstruction Examples

CIFAR-10

Sample From Test Set

Implicit GON

Super Resolution Examples

MNIST

Sample From Test Set Sample From Test Set Sample From Test Set

Fashion-MNIST

Sample From Test Set Sample From Test Set

Variational GON

Variational GON (ngf:16, nz = 256)

MNIST Test Set Reconstruction

Sample From Test Set

Fashion-MNIST Test Set Reconstruction

Sample From Test Set

MNIST Sampling

Sample From Test Set

Fashion-MNIST Sampling

Sample From Test Set

Ablation Studies

GON Activation Function

GONs require activation functions with non-zero second derivative. We trained 3 same GON architecture with ELU,RELU, and SELU on MNIST for 50 epochs as illustrated in the next figure. Note that GON with RELU underperforms. See Ablation_Studies folder for training experiments. Sample From Test Set

GON Latent Initialization

Initialization of latent at the origin is sufficient for GONs. We also experimented with random initialization of latent vector drawn from zero mean diagonal sigma square covariance matrix for various sigma (standard deviation) values. Note that as we increase the sigma, the GON model performs worse. Sample From Test Set

Refences

[1] Sam Bond-Taylor and Chris G. Willcocks. Gradient origin networks. In International Conference on Learning Representations, 2021. URL https://openreview.net/pdf?id=0O_cQfw6uEh.

[2] https://github.com/cwkx/GON

[3] Vincent Sitzmann, Julien N. P. Martel, Alexander W. Bergman, David B. Lindell, and Gordon Wet- zstein. Implicit neural representations with periodic activation functions, 2020.

About

Reimplementation of Gradient Origin Networks in Julia with Knet Framework

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages