CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-30 08:42:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama: fix missing k_cache store for rwkv6qwen2 (#11445)

Browse Source 
Signed-off-by: Molly Sophia <mollysophia379@gmail.com>
This commit is contained in:
Molly Sophia
2025-01-29 12:07:21 +08:00
committed by GitHub
parent 794fe23f29
commit 325afb370a
1 changed files with 3 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/llama.cpp
View File

@@ -7700,17 +7700,13 @@ struct llm_build_context {
1
);
struct ggml_tensor * last_norm_att = ggml_view_3d(ctx0, x_norm_att, n_embd, 1, n_seqs, x_norm_att->nb[1], x_norm_att->nb[2], (n_seq_tokens-1)*n_embd*ggml_element_size(x_norm_att));
ggml_build_forward_expand(
gf,
ggml_cpy(
ctx0,
wkv_states,
ggml_view_1d(
ctx0,
kv_self.v_l[il],
hparams.n_embd_v_s() * n_seqs,
hparams.n_embd_v_s() * kv_head * ggml_element_size(kv_self.v_l[il])
)
ggml_view_1d(ctx0, last_norm_att, n_embd * n_seqs, 0),
ggml_view_1d(ctx0, kv_self.k_l[il], hparams.n_embd_k_s() * n_seqs, hparams.n_embd_k_s() * kv_head * ggml_element_size(kv_self.k_l[il]))
)
);