HIP: broaden MoE ncols_opt tile heuristic on RDNA3.5 architecture ( #28935 ) It's found the MoE ncols_opt tile heuristic needs to be broadened to include the RDNA3.5 architecture. The code change is implemented in ggml/src/ggml-cuda/mmq.cu and just change the…
Read the original source — github.com
release · Shared by tscosj
0 comments
No comments yet.