llama.cpp
1ad42b1f - ggml : ggml_soft_max uses F16 mask

Commit
1 year ago
ggml : ggml_soft_max uses F16 mask
Author
Committer
Parents
Loading