Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-27 06:39:25 +01:00
Code Issues Packages Projects Releases Wiki Activity

llama : make pos_bias contiguous for CUDA

Browse Source
This commit is contained in:
Stanisław Szymczyk 2024-06-26 17:46:39 +02:00
parent bad0cafee9
commit c4ded1a8fb
1 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
llama.cpp
View File

@ -8104,6 +8104,9 @@ struct llm_build_context {
pos_bias = ggml_permute(ctx0, pos_bias, 2, 0, 1, 3); pos_bias = ggml_permute(ctx0, pos_bias, 2, 0, 1, 3);
cb(pos_bias, "pos_bias", -1); cb(pos_bias, "pos_bias", -1);
pos_bias = ggml_cont(ctx0, pos_bias);
cb(pos_bias, "pos_bias", -1);
return pos_bias; return pos_bias;
} }