Sync master with upstream release b6240#211

Merged

Minh141120 merged 30 commits intodevfrom

update-dev-from-master-2025-08-22-00-11

Aug 22, 2025

Commits on Aug 22, 2025

musa: fix build warnings (#15258)

yeahdongcn
authored and
qnixsynapse
committed
lookahead : add sample command to readme (#15447)

ggerganov
authored and
qnixsynapse
committed
cmake : fix target include directories (#15450)

xiaobing318
authored and
qnixsynapse
committed
common : fix context shift help message (#15448)

DamonFool
authored and
qnixsynapse
committed
chat: handle gpt-oss return/end token inconsistency (#15421)

danbev
authored and
qnixsynapse
committed
vulkan: shorten pipeline name strings (#15431)

jeffbolznv
authored and
qnixsynapse
committed
CUDA: replace GGML_CUDA_F16 with CUDA arch checks (#15433)

JohannesGaessler
authored and
qnixsynapse
committed
CUDA: refactor FA support/selection code (#15454)

JohannesGaessler
authored and
qnixsynapse
committed
server: fix OpenAI API compatibility for usage statistics in chat streams (#15444)

TeoZosa
authored and
qnixsynapse
committed
sched : copy only the used experts when offloading prompt processing (#15346)

slaren
authored and
qnixsynapse
committed
musa: add GGML_UNUSED_VARS (#15446)

yeahdongcn
authored and
qnixsynapse
committed
examples : remove references to `make` in examples [no ci] (#15457)

danbev
authored and
qnixsynapse
committed
server : fix webui (#15462)

stduhpf
authored and
qnixsynapse
committed
ggml : fix condition of im2col on Metal backend (#15460)

ngxson
authored and
qnixsynapse
committed
common : fix incorrect print of non-ascii characters in the logging (#15466)

DamonFool
authored and
qnixsynapse
committed
convert : make Mistral community chat templates optional via parameter (#15420)

juliendenize
authored and
qnixsynapse
committed
ci : add copilot-instructions.md (#15286)

authored and
qnixsynapse
committed
ci : fix -Werror=return-type in clip.cpp so ci/run.sh can run without issue (#15221)

authored and
qnixsynapse
committed
examples : add model conversion tool/example (#15455)

danbev
authored and
qnixsynapse
committed
ci : continue file download with wget (#15471)

ggerganov
authored and
qnixsynapse
committed
ci : enable RVV1.0 native build (#15386)

authored and
qnixsynapse
committed
examples : install torch-cpu for model conversion tool/example (#15475)

DamonFool
authored and
qnixsynapse
committed
kv-cache : drop the "unified" prefix (#15467)

ggerganov
authored and
qnixsynapse
committed
examples : fix some typos in examples/model-conversion/README.md (#15477)

DamonFool
authored and
qnixsynapse
committed
vulkan: Reuse conversion results in prealloc_y (#15410)

jeffbolznv
authored and
qnixsynapse
committed
vulkan: add exp operation (#15456)

authored and
qnixsynapse
committed
vulkan : support conv_2d_dw with f16 weights (#15392)

Acly
authored and
qnixsynapse
committed
graph : remove build_attn_with_sinks overload (#15469)

ggerganov
authored and
qnixsynapse
committed
llama : remove deprecated llama_kv_self API (#15472)

ggerganov
authored and
qnixsynapse
committed
sched : fix possible use of wrong ids tensor when offloading moe prompt processing (#15488)

slaren
authored and
qnixsynapse
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Sync master with upstream release b6240#211

Sync master with upstream release b6240#211
Minh141120 merged 30 commits intodevfrom
update-dev-from-master-2025-08-22-00-11

Commits on Aug 22, 2025

musa: fix build warnings (#15258)

lookahead : add sample command to readme (#15447)

cmake : fix target include directories (#15450)

common : fix context shift help message (#15448)

chat: handle gpt-oss return/end token inconsistency (#15421)

vulkan: shorten pipeline name strings (#15431)

CUDA: replace GGML_CUDA_F16 with CUDA arch checks (#15433)

CUDA: refactor FA support/selection code (#15454)

server: fix OpenAI API compatibility for usage statistics in chat streams (#15444)

sched : copy only the used experts when offloading prompt processing (#15346)

musa: add GGML_UNUSED_VARS (#15446)

examples : remove references to `make` in examples [no ci] (#15457)

server : fix webui (#15462)

ggml : fix condition of im2col on Metal backend (#15460)

common : fix incorrect print of non-ascii characters in the logging (#15466)

convert : make Mistral community chat templates optional via parameter (#15420)

ci : add copilot-instructions.md (#15286)

ci : fix -Werror=return-type in clip.cpp so ci/run.sh can run without issue (#15221)

examples : add model conversion tool/example (#15455)

ci : continue file download with wget (#15471)

ci : enable RVV1.0 native build (#15386)

examples : install torch-cpu for model conversion tool/example (#15475)

kv-cache : drop the "unified" prefix (#15467)

examples : fix some typos in examples/model-conversion/README.md (#15477)

vulkan: Reuse conversion results in prealloc_y (#15410)

vulkan: add exp operation (#15456)

vulkan : support conv_2d_dw with f16 weights (#15392)

graph : remove build_attn_with_sinks overload (#15469)

llama : remove deprecated llama_kv_self API (#15472)

sched : fix possible use of wrong ids tensor when offloading moe prompt processing (#15488)

Sync master with upstream release b6240#211Minh141120 merged 30 commits intodevjanhq/llama.cpp:devfrom update-dev-from-master-2025-08-22-00-11janhq/llama.cpp:update-dev-from-master-2025-08-22-00-11Copy head branch name to clipboard

Commits

Commits on Aug 22, 2025

Sync master with upstream release b6240#211
Minh141120 merged 30 commits intodevfrom
update-dev-from-master-2025-08-22-00-11