Merge commit '79bba02a6741de194912d370015866414faa83ad' into concedo_experimental

# Conflicts:
#	.github/workflows/build-self-hosted.yml
#	.github/workflows/release.yml
#	.github/workflows/server-self-hosted.yml
#	ggml/src/ggml-opencl/ggml-opencl.cpp
#	ggml/src/ggml-opencl/kernels/flash_attn_f32_f16.cl
#	ggml/src/ggml-opencl/kernels/flash_attn_f32_q4_0.cl
#	ggml/src/ggml-opencl/kernels/flash_attn_f32_q8_0.cl
#	ggml/src/ggml-opencl/kernels/gemm_moe_mxfp4_f32_ns.cl
#	ggml/src/ggml-opencl/kernels/mul_mv_f16_f32_l4.cl
#	ggml/src/ggml-opencl/kernels/mul_mv_q4_k_f32.cl
#	tests/test-backend-ops.cpp
#	tools/tokenize/tokenize.cpp
#	tools/ui/src/lib/components/app/chat/ChatForm/ChatFormActions/ChatFormActionAdd/ChatFormActionAddToolsSubmenu.svelte
This commit is contained in:
Concedo
2026-07-18 11:04:46 +08:00
180 changed files with 7455 additions and 1324 deletions
+4 -1
View File
@@ -260,7 +260,10 @@ struct common_speculative_impl_draft_simple : public common_speculative_impl {
bool process(const llama_batch & batch) override {
auto * ctx_dft = params.ctx_dft;
const int ret = llama_decode(ctx_dft, batch);
llama_batch batch_dft = batch;
batch_dft.logits = nullptr;
const int ret = llama_decode(ctx_dft, batch_dft);
if (ret != 0) {
SPC_ERR("failed to decode draft batch, ret = %d\n", ret);