vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 09:34:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'upstream' into concedo_experimental

Browse source 
# Conflicts:
#	.github/workflows/build.yml
#	.github/workflows/close-issue.yml
#	.github/workflows/server.yml
#	AUTHORS
#	CMakeLists.txt
#	Makefile
#	README.md
#	cmake/llama.pc.in
#	common/CMakeLists.txt
#	docs/build.md
#	examples/batched.swift/Sources/main.swift
#	examples/llama.swiftui/llama.cpp.swift/LibLlama.swift
#	examples/llava/CMakeLists.txt
#	examples/llava/clip.h
#	examples/run/run.cpp
#	examples/server/README.md
#	ggml/CMakeLists.txt
#	ggml/src/ggml-cuda/CMakeLists.txt
#	ggml/src/ggml-hip/CMakeLists.txt
#	ggml/src/ggml-musa/CMakeLists.txt
#	scripts/sync-ggml.last
#	tests/CMakeLists.txt
#	tests/test-backend-ops.cpp
#	tests/test-chat-template.cpp
#	tests/test-grammar-integration.cpp
#	tests/test-json-schema-to-grammar.cpp
This commit is contained in:
Concedo 2025-02-07 00:52:31 +08:00
commit db6db9dff9
142 changed files with 74261 additions and 10014 deletions
Download patch file Download diff file Expand all files Collapse all files

41
src/llama.cpp
View file

@ -4646,7 +4646,8 @@ struct llm_build_context {
ggml_row_size(kv_pe_compresseed->type, kv_lora_rank));
cb(k_pe, "k_pe", il);
kv_compressed = ggml_cont(ctx0, kv_compressed); // TODO: the CUDA backend does not support non-contiguous norm
// TODO: the CUDA backend used to not support non-cont. (RMS) norm, investigate removing ggml_cont
kv_compressed = ggml_cont(ctx0, kv_compressed);
kv_compressed = llm_build_norm(ctx0, kv_compressed, hparams,
model.layers[il].attn_kv_a_norm, NULL,
LLM_NORM_RMS, cb, il);
@ -6500,7 +6501,8 @@ struct llm_build_context {
ggml_row_size(kv_pe_compresseed->type, kv_lora_rank));
cb(k_pe, "k_pe", il);
kv_compressed = ggml_cont(ctx0, kv_compressed); // TODO: the CUDA backend does not support non-contiguous norm
// TODO: the CUDA backend used to not support non-cont. (RMS) norm, investigate removing ggml_cont
kv_compressed = ggml_cont(ctx0, kv_compressed);
kv_compressed = llm_build_norm(ctx0, kv_compressed, hparams,
model.layers[il].attn_kv_a_norm, NULL,
LLM_NORM_RMS, cb, il);
@ -7251,17 +7253,30 @@ struct llm_build_context {
struct ggml_tensor * Qcur = nullptr;
struct ggml_tensor * Kcur = nullptr;
struct ggml_tensor * Vcur = nullptr;
cur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wqkv, cur);
cb(cur, "wqkv", il);
cur = ggml_add(ctx0, cur, model.layers[il].bqkv);
cb(cur, "bqkv", il);
Qcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd, n_tokens, cur->nb[1], 0*sizeof(float)*(n_embd)));
Kcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd_gqa, n_tokens, cur->nb[1], 1*sizeof(float)*(n_embd)));
Vcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd_gqa, n_tokens, cur->nb[1], 1*sizeof(float)*(n_embd + n_embd_gqa)));
if (model.type == LLM_TYPE_1_5B || model.type == LLM_TYPE_4B || model.type == LLM_TYPE_9B) {
Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
if (model.layers[il].bq) {
Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
}
Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
if (model.layers[il].bk) {
Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
}
Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
if (model.layers[il].bv) {
Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
}
} else {
cur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wqkv, cur);
cb(cur, "wqkv", il);
if (model.layers[il].bqkv) {
cur = ggml_add(ctx0, cur, model.layers[il].bqkv);
cb(cur, "bqkv", il);
}
Qcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd, n_tokens, cur->nb[1], 0*sizeof(float)*(n_embd)));
Kcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd_gqa, n_tokens, cur->nb[1], 1*sizeof(float)*(n_embd)));
Vcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd_gqa, n_tokens, cur->nb[1], 1*sizeof(float)*(n_embd + n_embd_gqa)));
}
cb(Qcur, "Qcur", il);
cb(Kcur, "Kcur", il);
cb(Vcur, "Vcur", il);