happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

removed Kimi-Linear specific change that causes failure at server-windows

This commit is contained in:
Yee Man Chan 2026-01-13 16:31:29 +08:00
parent 217e7ce45a
commit 6ba78d1220
1 changed files with 5 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/llama-graph.cpp
View File

@ -1963,15 +1963,11 @@ ggml_tensor * llm_graph_context::build_rs(
ggml_tensor * output_states = get_state_rows(ctx0, states, state_copy_main); ggml_tensor * output_states = get_state_rows(ctx0, states, state_copy_main);
ggml_build_forward_expand(gf, output_states); ggml_build_forward_expand(gf, output_states);
// copy extra states which won't be changed further (between n_seqs and n_rs) ggml_tensor * states_extra = ggml_get_rows(ctx0, states, state_copy_extra);
// Skip if there are no extra states to copy (n_rs == n_seqs) ggml_build_forward_expand(gf,
if (arch != LLM_ARCH_KIMI_LINEAR || n_rs > (u_int32_t) n_seqs) { // arch check for backward compat ggml_cpy(ctx0,
ggml_tensor * states_extra = ggml_get_rows(ctx0, states, state_copy_extra); states_extra,
ggml_build_forward_expand(gf, ggml_view_1d(ctx0, s, state_size*(n_rs - n_seqs), (rs_head + n_seqs)*state_size*ggml_element_size(s))));
ggml_cpy(ctx0,
states_extra,
ggml_view_1d(ctx0, s, state_size*(n_rs - n_seqs), (rs_head + n_seqs)*state_size*ggml_element_size(s))));
}
return output_states; return output_states;
} }