Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 09:15:18 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge commit 'c96f608d9861f7e8466bc1b6ac2ff4e3c6f96641' into concedo_experimental

Browse Source 
# Conflicts:
#	CONTRIBUTING.md
#	docs/ops.md
#	docs/ops/Vulkan.csv
#	models/templates/LFM2-8B-A1B.jinja
#	tests/peg-parser/test-python-dict-parser.cpp
#	tests/peg-parser/test-unicode.cpp
#	tests/test-chat-peg-parser.cpp
#	tests/test-chat.cpp
#	tools/llama-bench/llama-bench.cpp
This commit is contained in:
Concedo
2026-03-14 17:14:34 +08:00
parent 8b9594b6ea c96f608d98
commit ff3f8533d3
142 changed files with 1231 additions and 635 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-context.cpp
+12 -8
View File
@@ -2886,19 +2886,23 @@ llama_context * llama_init_from_model(
if (params.flash_attn_type == LLAMA_FLASH_ATTN_TYPE_AUTO && ggml_is_quantized(params.type_k)) {
const uint32_t blck_size = ggml_blck_size(params.type_k);
if (model->hparams.n_embd_head_k % blck_size != 0) {
LLAMA_LOG_ERROR("%s: K cache type %s with block size %u does not divide n_embd_head_k=%u\n",
__func__, ggml_type_name(params.type_k), blck_size, model->hparams.n_embd_head_k);
return nullptr;
for (uint32_t il = 0; il < model->hparams.n_layer; ++il) {
if (model->hparams.n_embd_head_k(il) % blck_size != 0) {
LLAMA_LOG_ERROR("%s: K cache type %s with block size %u does not divide n_embd_head_k=%u\n",
__func__, ggml_type_name(params.type_k), blck_size, model->hparams.n_embd_head_k(il));
return nullptr;
}
}
}
if (params.flash_attn_type == LLAMA_FLASH_ATTN_TYPE_AUTO && ggml_is_quantized(params.type_v)) {
const uint32_t blck_size = ggml_blck_size(params.type_v);
if (model->hparams.n_embd_head_v % blck_size != 0) {
LLAMA_LOG_ERROR("%s: V cache type %s with block size %u does not divide n_embd_head_k=%u\n",
__func__, ggml_type_name(params.type_v), blck_size, model->hparams.n_embd_head_v);
return nullptr;
for (uint32_t il = 0; il < model->hparams.n_layer; ++il) {
if (model->hparams.n_embd_head_v(il) % blck_size != 0) {
LLAMA_LOG_ERROR("%s: V cache type %s with block size %u does not divide n_embd_head_v=%u\n",
__func__, ggml_type_name(params.type_v), blck_size, model->hparams.n_embd_head_v(il));
return nullptr;
}
}
}