One way to integrate into existing CUDA infrastructure

FluxML / Torch.jl

Sensible extensions for exposing torch in Julia.

Other

213 stars 15 forks source link

One way to integrate into existing CUDA infrastructure #2

Open DhairyaLGandhi opened 4 years ago

DhairyaLGandhi commented 4 years ago

This seems to work just fine, and we can move between CuArrays and Tensors fairly cheaply. If this is enough to do the trick, we should move this into https://github.com/dhairyagandhi96/TorchGPU.jl

DhairyaLGandhi commented 4 years ago

bors ping

bors[bot] commented 4 years ago

pong