From: Georgi Gerganov <redacted>
Date: Thu, 18 Sep 2025 09:47:56 +0000 (+0300)
Subject: llama : bump max seq limit from 64 to 256 (#15916)
X-Git-Tag: upstream/0.0.6527~20
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=e58174cecbc45bf79bf653cd2c984395940c6ef4;p=pkg%2Fggml%2Fsources%2Fllama.cpp

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

ggml-ci
---

diff --git a/src/llama-cparams.h b/src/llama-cparams.h
index dbbaba9f..eae7b839 100644
--- 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