From 96d7e48bd0ae359e2ffe165807083c775093d7a6 Mon Sep 17 00:00:00 2001 From: "L. Schuster" <laurin.schuster@fau.de> Date: Thu, 7 Apr 2022 21:02:43 +0100 Subject: [PATCH] fixed typo --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index a98d608..a73d664 100644 --- a/README.md +++ b/README.md @@ -106,7 +106,7 @@ ACGAN first output without improvements (CGAN with multilabel): | :heavy_check_mark: | Discriminator learns too fast | Use more class labels to make training harder (called ACGAN) | Slower convergence of the discriminator | No stable equilibrium | | :heavy_check_mark: | Generator not able to produce alike images | Increase the parameter size of the UNet | Noteable faster convergence of the generator | No stable equilibrium | | :heavy_check_mark: | Discriminator fooled by wrong images | Implement lager discriminator with ResNet-18 adaption | Noteable faster convergence of the discriminator | No stable equilibrium | -| :heavy_multiplication_x: | BCE Loss not favored with GANs | Using L2-Loss | Installable loss but longer convergence | No stable equilibrium | +| :heavy_multiplication_x: | BCE Loss not favored with GANs | Using L2-Loss | Unstable loss but longer convergence | No stable equilibrium | Best ACGAN with all improvements: -- GitLab