llama : bump max seq limit from 64 to 256 (#15916)

ggml-ci
2025-09-18 12:47:56 +03:00 · 2025-09-18 12:47:56 +03:00 · e58174cecb
parent b213fce89b
commit e58174cecb
1 changed files with 1 additions and 1 deletions
--- a/src/llama-cparams.h
+++ b/src/llama-cparams.h
@ -4,7 +4,7 @@

 #include <cstdint>

-#define LLAMA_MAX_SEQ 64
+#define LLAMA_MAX_SEQ 256

 struct llama_cparams {
    uint32_t n_ctx;           // context size used during inference