issues
search
hikettei
/
cl-waffe2
[Experimental] Graph and Tensor Abstraction for Deep Learning all in Common Lisp
https://hikettei.github.io/cl-waffe2/
MIT License
142
stars
6
forks
source link
[Optimize] 2x times faster than PyTorch!!!!
#119
Closed
hikettei
closed
1 year ago
hikettei
commented
1 year ago
[Optimize] Enabled Loop Collapse anywhere
[Optimize] Parameters are collapsed well
As a result, Training MLP model is now 2x times faster than pure PyTorch!!!