issues
search
replicate
/
cog-triton
A cog implementation of Nvidia's Triton server
Apache License 2.0
11
stars
0
forks
source link
Merge nvidia-*-cu12 python with nix's cudaPackages
#31
Closed
yorickvP
closed
4 months ago
yorickvP
commented
4 months ago
These are actually the same redistributables, so just symlink them. Saves a bunch of space. The images are now 5.4GB.
Remove a bunch of cuda kernel source files.
Also fix that annoying
libtritonserver-90a4cf82.so
patch.
Tested and working. Pushed to cache:
[x] runner-86
[x] runner-80
[x] runner-90
[x] builder
libtritonserver-90a4cf82.sopatch.Tested and working. Pushed to cache: