Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-20 01:31:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'upstream' into concedo_experimental

Browse Source 
# Conflicts:
#	.devops/intel.Dockerfile
#	.github/workflows/python-type-check.yml
#	embd_res/templates/Qwen3.5-4B.jinja
#	examples/model-conversion/scripts/causal/compare-logits.py
#	examples/model-conversion/scripts/utils/check-nmse.py
#	examples/model-conversion/scripts/utils/compare_tokens.py
#	examples/model-conversion/scripts/utils/semantic_check.py
#	examples/sycl/build.sh
#	examples/sycl/run-llama2.sh
#	ggml/src/ggml-hexagon/htp/flash-attn-ops.c
#	ggml/src/ggml-hexagon/htp/hex-dma.h
#	ggml/src/ggml-hexagon/htp/rope-ops.c
#	scripts/gen-unicode-data.py
#	tests/test-chat.cpp
This commit is contained in:
Concedo
2026-03-30 21:41:19 +08:00
parent 9864d46389 389c7d4955
commit a3a5897d93
20 changed files with 565 additions and 152 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-arch.cpp
+2
View File
@@ -557,6 +557,8 @@ static std::set<llm_tensor> llm_get_tensor_names(llm_arch arch) {
LLM_TENSOR_OUTPUT_NORM,
LLM_TENSOR_OUTPUT,
LLM_TENSOR_ROPE_FREQS,
LLM_TENSOR_ROPE_FACTORS_LONG,
LLM_TENSOR_ROPE_FACTORS_SHORT,
LLM_TENSOR_ATTN_NORM,
LLM_TENSOR_ATTN_Q,
LLM_TENSOR_ATTN_K,
src/llama-model-loader.cpp
+6
View File
@@ -1162,6 +1162,12 @@ struct ggml_tensor * llama_model_loader::create_tensor(
if (overrides->buft == ggml_backend_cpu_buffer_type()) {
// when overriding to a CPU buffer, consider the extra buffer types
buft = select_weight_buft(hparams, t_meta, op, buft_list_cpu);
if (use_mmap) {
static std::once_flag once;
std::call_once(once, [] {
LLAMA_LOG_WARN("llama_model_loader: tensor overrides to CPU are used with mmap enabled - consider using --no-mmap for better performance\n");
});
}
} else {
buft = overrides->buft;
}