Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

llama-batch: add n_keep_tail in split_equal for recurrent models (#25278)

Browse Source
This commit is contained in:
Aman Gupta
2026-07-08 15:55:19 +08:00
committed by GitHub
parent f296fdfbed
commit 230ea9d214
10 changed files with 102 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-kv-cache.cpp
+1 -1
View File
@@ -706,7 +706,7 @@ llama_memory_context_ptr llama_kv_cache::init_batch(
std::vector<llama_ubatch> ubatches;
while (true) {
auto ubatch = n_stream == 1 ? balloc.split_simple(n_ubatch) : balloc.split_equal(n_ubatch, true);
auto ubatch = n_stream == 1 ? balloc.split_simple(n_ubatch) : balloc.split_equal(n_ubatch, true, 0);
if (ubatch.n_tokens == 0) {
break;