Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-17 08:19:49 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge commit '157b81fe6dbfec7d7ce91ef7cd9c6bc0c218d6fe' into concedo_experimental

Browse Source 
# Conflicts:
#	.devops/rocm.Dockerfile
#	.github/workflows/build-cuda-ubuntu.yml
#	.github/workflows/build-cuda-windows.yml
#	.github/workflows/build-sanitize.yml
#	.github/workflows/release.yml
#	README.md
#	ci/run.sh
#	ggml/src/ggml-webgpu/ggml-webgpu-shader-lib.hpp
#	ggml/src/ggml-webgpu/ggml-webgpu.cpp
#	ggml/src/ggml-webgpu/wgsl-shaders/concat.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/conv2d.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/conv2d_dw.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/flash_attn.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_tile.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_split.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/im2col.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/rms_norm_mul.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/row_norm.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/soft_max.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/solve_tri.wgsl
#	ggml/src/ggml-webgpu/wgsl-shaders/ssm_scan.wgsl
#	tests/test-backend-ops.cpp
#	tests/test-llama-archs.cpp
#	tools/cli/README.md
#	tools/completion/README.md
#	tools/server/README.md
#	tools/ui/src/lib/constants/settings-registry.ts
#	tools/ui/src/lib/hooks/use-models-selector.svelte.ts
#	tools/ui/src/lib/hooks/use-tools-panel.svelte.ts
#	tools/ui/src/lib/services/chat.service.ts
#	tools/ui/svelte.config.js
#	tools/ui/tests/stories/a11y/ChatScreenForm.a11y.stories.svelte
This commit is contained in:
Concedo
2026-08-10 21:05:07 +08:00
parent 59d92956f7 157b81fe6d
commit 910500e78f
566 changed files with 10579 additions and 6163 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-context.cpp
+2 -1
View File
@@ -3612,8 +3612,9 @@ llama_context * llama_init_from_model(
model->hparams.pooling_type, params.pooling_type);
}
// router_layer >= 0 means n_layer_nextn is repurposed for a router layer, not real MTP
if (params.ctx_type == LLAMA_CONTEXT_TYPE_MTP &&
model->hparams.n_layer_nextn == 0) {
(model->hparams.n_layer_nextn == 0 || model->hparams.router_layer >= 0)) {
LLAMA_LOG_WARN("%s: context type MTP requested but model doesn't contain MTP layers\n", __func__);
return nullptr;
}