- Category
- Developer Tools
- Rank
- No. 3386Tools index
- Type
- TOOL
- GitHub
- 2 stars
- Latest release
- v0.1
- Date
About
CNet is a C++/CUDA framework for building and researching deep complex-valued neural networks and for optimizing complex functions by gradient descent with Wirtinger derivatives. Each built-in layer has a CPU reference and a CUDA kernel, and it includes Fourier-transform and Hadamard layers plus a Born-rule loss. It targets research on complex-valued networks rather than competing with mainstream frameworks.
Tags
neural-networkscomplex-valuedcudacppdeep-learningresearch
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.