tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-05 08:28:37 -04:00
Code Issues Packages Projects Releases Wiki Activity

llama : fix Gemma3 SWA KV cache shift (#12373)

Browse Source 
* llama : fix Gemma3 SWA KV cache shift

ggml-ci

* hparams : add comment [no ci]
This commit is contained in:
Georgi Gerganov
2025-03-13 19:08:07 +02:00
committed by GitHub
parent be7c303410
commit 84d5475541
6 changed files with 37 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/llama-hparams.h
View File

@@ -36,6 +36,7 @@ struct llama_hparams {
uint32_t n_layer;
uint32_t n_rot;
uint32_t n_swa = 0; // sliding window attention (SWA)
uint32_t n_swa_pattern = 1; // by default, all layers use non-sliding-window attention
uint32_t n_embd_head_k; // dimension of keys (d_k). d_q is assumed to be the same, but there are n_head q heads, and only n_head_kv k-v heads
uint32_t n_embd_head_v; // dimension of values (d_v) aka n_embd_head
uint32_t n_expert = 0;
@@ -133,6 +134,8 @@ struct llama_hparams {
// dimension of the recurrent state embeddings
uint32_t n_embd_v_s() const;
bool is_sliding(uint32_t il) const;
};
static_assert(std::is_trivially_copyable<llama_hparams>::value, "llama_hparams must be trivially copyable");