Skip to content

Latest commit

 

History

History
5 lines (3 loc) · 285 Bytes

README.md

File metadata and controls

5 lines (3 loc) · 285 Bytes

umap-pca

B3 set uses the channel with the same size of the input shape in the convolution layer. For example, (1,256,120) for dataset with 120 components, (1,256,90) for dataset with 90 components to avoid overfitting problem.

B4 set use the exact same channel with the benchmark.