Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 01:05:09 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'upstream' into concedo_experimental

Browse Source 
# Conflicts:
#	.github/workflows/ui-build.yml
#	.github/workflows/ui-publish.yml
#	.github/workflows/ui-self-hosted.yml
#	CMakeLists.txt
#	app/CMakeLists.txt
#	app/llama.cpp
#	common/arg.cpp
#	ggml/CMakeLists.txt
#	ggml/src/ggml-hexagon/ggml-hexagon.cpp
#	ggml/src/ggml-hexagon/htp/htp-ops.h
#	ggml/src/ggml-hexagon/htp/main.c
#	ggml/src/ggml-hexagon/htp/unary-ops.c
#	ggml/src/ggml-opencl/ggml-opencl.cpp
#	scripts/snapdragon/ggml-hexagon-profile.py
#	scripts/sync-ggml.last
#	src/CMakeLists.txt
#	tests/test-llama-archs.cpp
#	tools/cli/README.md
#	tools/completion/README.md
#	tools/mtmd/CMakeLists.txt
#	tools/mtmd/tests/test-deepseek-ocr.py
#	tools/server/README.md
This commit is contained in:
Concedo
2026-05-30 01:54:53 +08:00
parent baacd640c0 06d26dfdff
commit e541554d70
55 changed files with 2443 additions and 251 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-model.cpp
+25 -1
View File
@@ -10,6 +10,7 @@
#include "llama-kv-cache.h"
#include "llama-kv-cache-iswa.h"
#include "llama-kv-cache-dsa.h"
#include "llama-memory-hybrid.h"
#include "llama-memory-hybrid-iswa.h"
#include "llama-memory-recurrent.h"
@@ -57,6 +58,7 @@
#include "models/deepseek.cpp"
#include "models/deepseek2.cpp"
#include "models/deepseek2ocr.cpp"
#include "models/deepseek32.cpp"
#include "models/delta-net-base.cpp"
#include "models/dots1.cpp"
#include "models/dream.cpp"
@@ -303,6 +305,8 @@ static llama_model * llama_model_mapping(llm_arch arch, const llama_model_params
return new llama_model_deepseek2(params);
case LLM_ARCH_DEEPSEEK2OCR:
return new llama_model_deepseek2ocr(params);
case LLM_ARCH_DEEPSEEK32:
return new llama_model_deepseek32(params);
case LLM_ARCH_GLM_DSA:
return new llama_model_glm_dsa(params);
case LLM_ARCH_MISTRAL4:
@@ -910,6 +914,7 @@ const char * llm_type_name(llm_type type) {
case LLM_TYPE_310B_A15B: return "310B.A15B";
case LLM_TYPE_355B_A32B: return "355B.A32B";
case LLM_TYPE_397B_A17B: return "397B.A17B";
case LLM_TYPE_685B_A37B: return "685B.A37B";
case LLM_TYPE_744B_A40B: return "744B.A40B";
case LLM_TYPE_E2B: return "E2B";
case LLM_TYPE_E4B: return "E4B";
@@ -1901,7 +1906,7 @@ void llama_model::print_info() const {
LLAMA_LOG_INFO("%s: expert_weights_scale = %.1f\n", __func__, hparams.expert_weights_scale);
}
if (arch == LLM_ARCH_DEEPSEEK2 || arch == LLM_ARCH_DEEPSEEK2OCR || arch == LLM_ARCH_GLM_DSA || arch == LLM_ARCH_MISTRAL4) {
if (arch == LLM_ARCH_DEEPSEEK2 || arch == LLM_ARCH_DEEPSEEK2OCR || arch == LLM_ARCH_DEEPSEEK32 || arch == LLM_ARCH_GLM_DSA || arch == LLM_ARCH_MISTRAL4) {
LLAMA_LOG_INFO("%s: n_layer_dense_lead = %d\n", __func__, hparams.n_layer_dense_lead);
LLAMA_LOG_INFO("%s: n_lora_q = %d\n", __func__, hparams.n_lora_q);
LLAMA_LOG_INFO("%s: n_lora_kv = %d\n", __func__, hparams.n_lora_kv);
@@ -2089,6 +2094,23 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
{
res = nullptr;
} break;
case LLM_ARCH_DEEPSEEK32:
{
res = new llama_kv_cache_dsa(
*this,
params.type_k,
params.type_v,
!cparams.flash_attn,
cparams.offload_kqv,
cparams.kv_unified,
cparams.n_ctx_seq,
cparams.n_seq_max,
1,
hparams.n_swa,
hparams.swa_type,
nullptr,
nullptr);
} break;
// Models that need standard caching should rely on recurrent/hybrid
// checks
default:
@@ -2215,6 +2237,7 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
res = new llama_kv_cache(
*this,
hparams,
params.type_k,
params.type_v,
!cparams.flash_attn,
@@ -2404,6 +2427,7 @@ llama_rope_type llama_model_rope_type(const llama_model * model) {
case LLM_ARCH_DEEPSEEK:
case LLM_ARCH_DEEPSEEK2:
case LLM_ARCH_DEEPSEEK2OCR:
case LLM_ARCH_DEEPSEEK32:
case LLM_ARCH_PLM:
case LLM_ARCH_CHATGLM:
case LLM_ARCH_GRANITE: