update readme

This commit is contained in:
Harry Stuart 2024-06-08 11:03:06 +10:00
parent 4dcee06b0a
commit 6e96b37955

View File

@ -1,10 +1,10 @@
# Generative Adverserial Network for generating images # Generative Adverserial Network for generating images
The code here is a slightly altered version of the code that has been generously made publicly available at https://github.com/gsurma/image_generator The code here is an altered version of the code that has been generously made publicly available at https://github.com/gsurma/image_generator
As an introduction to the world of GANs, I decided to try the above model on some different datasets. You can see the results of two example trainings in the respective output folders. As an introduction to the world of GANs, I decided to try the above model on some different datasets. You can see the results of two example trainings in the respective output folders.
Training on human faces for 100 epochs took around 13 hours using the hardware mentioned on my profile. I am yet to optimise the architecture and hyperparameters. Training on human faces for 100 epochs took around 13 hours on an RTX 2060.
![samples_183](https://user-images.githubusercontent.com/42882697/206129682-e972310f-7aba-427c-b7fd-12ee46f40fc9.png) ![samples_183](https://user-images.githubusercontent.com/42882697/206129682-e972310f-7aba-427c-b7fd-12ee46f40fc9.png)