Closed wkcn closed 6 years ago
In current implementation, function dot and im2col are very slow.
The naive dot has low cache hit.
I will find the bottle necks and fix them.
Test Code
In the version db016f2.
The reason is that the single implementation couldn't perform well in all contexts. Meanwhile, the built-in Fully Connected Layer and Convolutional Layer have been discarded.
Fully Connected Layer
Convolutional Layer
In current implementation, function dot and im2col are very slow.
The naive dot has low cache hit.
I will find the bottle necks and fix them.
Test Code
In the version db016f2.