cbc385ca10f9be79f742db800b174c5dbb4e7bab
save training dataset by it target type. while training NN use single target training set to backward NN. this improve at least 20 times faster than last update!
Description
Languages
C#
77.3%
Jupyter Notebook
11.6%
Python
7.3%
ShaderLab
1.9%
GLSL
1.3%
Other
0.6%