Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 17:25:07 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'upstream' into concedo_experimental

Browse Source 
# Conflicts:
#	.github/workflows/build.yml
#	Makefile
#	docs/build.md
#	examples/rpc/rpc-server.cpp
#	examples/sycl/build.sh
#	ggml/CMakeLists.txt
#	ggml/src/ggml-cann/aclnn_ops.cpp
#	ggml/src/ggml-cann/ggml-cann.cpp
#	ggml/src/ggml-hip/CMakeLists.txt
#	scripts/sync-ggml.last
This commit is contained in:
Concedo
2025-04-17 00:52:37 +08:00
parent a5c39143a1 b43d89e311
commit 06159939d9
72 changed files with 6549 additions and 5397 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-kv-cache.cpp
+1 -1
View File
@@ -27,7 +27,7 @@ bool llama_kv_cache_unified::init(
recurrent = llama_model_is_recurrent(&model);
v_trans = !recurrent && !cparams.flash_attn;
can_shift = !recurrent && model.arch != LLM_ARCH_DEEPSEEK2; // not supported due to MLA
can_shift = !recurrent;
LLAMA_LOG_INFO("%s: kv_size = %d, offload = %d, type_k = '%s', type_v = '%s', n_layer = %d, can_shift = %d\n",
__func__, kv_size, offload, ggml_type_name(type_k), ggml_type_name(type_v), n_layer, can_shift);