llama : remove KV cache defragmentation logic (#15473)

ggml-ci
This commit is contained in:
Georgi Gerganov authored and GitHub committed 2025-08-22 12:22:13 +03:00
1 parent ad5c975c2d
commit 9ebebef62f
16 files changed
+32 -440

No files matched your search

+1 -1
View File
@@ -312,7 +312,7 @@ extern "C" {
float yarn_beta_fast; // YaRN low correction dim
float yarn_beta_slow; // YaRN high correction dim
uint32_t yarn_orig_ctx; // YaRN original context size
float defrag_thold; // defragment the KV cache if holes/size > thold, <= 0 disabled (default)
float defrag_thold; // [DEPRECATED] defragment the KV cache if holes/size > thold, <= 0 disabled (default)
ggml_backend_sched_eval_callback cb_eval;
void * cb_eval_user_data;