Name		Name	Last commit message	Last commit date
parent directory ..
gen_img		gen_img
infogan_model.py		infogan_model.py
infogan_tb.png		infogan_tb.png
infogan_train.py		infogan_train.py
readme.md		readme.md

readme.md

Information Maximizing Generative Adversarial Networks

Loss Function

used GAN loss at D/G nets.

Architecture Networks

Same with the InfoGAN paper.

DIFFS	InfoGAN Paper	ME
G fc units	`2 * 2 * 448`	`2 * 2 * 512`

Tensorboard

Elapsed time : s with GTX 1080 8GB x 1

Result

Name	Global Step 20k	Global Step 50k	Global Step 100k
InfoGAN

To-Do

MemoryError... I dunno why :(