]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
llama : use Q4_K for attn_v for Q2_K_S when n_gqa >= 4 (#4996)
authorKawrakow <redacted>
Wed, 17 Jan 2024 10:36:37 +0000 (12:36 +0200)
committerGitHub <redacted>
Wed, 17 Jan 2024 10:36:37 +0000 (12:36 +0200)
commit2b3a665d3917edf393761a24c4835447894df74a
tree254942a7222314ac60c406842912ae092d724000
parent75632936659772d5b2ce54b0b65319fecbaac2e6
llama : use Q4_K for attn_v for Q2_K_S when n_gqa >= 4 (#4996)

Co-authored-by: Iwan Kawrakow <redacted>
llama.cpp