sampling: Port of Smooth Sampling / Quadratic Sampling support by Silver267 · Pull Request #13441 · ggml-org/llama.cpp

Commits on May 10, 2025

Commits on May 11, 2025

Commits on May 13, 2025

Merge branch 'ggml-org:master' into master
Silver267
authored

Commits on May 14, 2025

vulkan: workaround FA compile failures on macos (#13517)

jeffbolznv
authored and
Silver267
committed
scripts : fix compare-llama-bench.py show parameter (#13514)

CISC
authored and
Silver267
committed
docs: Update link to ggml-org in multimodal.md (#13513)

authored and
Silver267
committed
webui: Allow pasting file from clipboard (#13526)

authored and
Silver267
committed
webui : use fflate for more deterministic gzip compress (#13525)

ngxson
authored and
Silver267
committed
vulkan: KHR_coopmat flash attention (#13506)

jeffbolznv
authored and
Silver267
committed
cmake: simplify vulkan shader test logic (#13263)

bandoti
authored and
Silver267
committed
server : fix cache_tokens bug with no cache_prompt (#13533)

ngxson
authored and
Silver267
committed
server : passthrough the /models endpoint during loading (#13535)

ggerganov
authored and
Silver267
committed
fix: Move build_inp_pos to the top of the graph section for build_granite (#13538)

gabe-l-hart
authored and
Silver267
committed
CUDA: faster Deepseek FA, add Turing support (#13435)

JohannesGaessler
authored and
Silver267
committed
llama : fix quantize with dl backends (#13539)

slaren
authored and
Silver267
committed
CUDA: fix crash on large batch size for quant. MoE (#13537)

JohannesGaessler
authored and
Silver267
committed
fix: crash when calling `llama_state_get_size` on a context without a KV cache (#13542)

giladgd
authored and
Silver267
committed
editorconfig : fix trailing whitespace from #13542 (#13546)

CISC
authored and
Silver267
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

sampling: Port of Smooth Sampling / Quadratic Sampling support#13441

sampling: Port of Smooth Sampling / Quadratic Sampling support#13441
Silver267 wants to merge 24 commits intoggml-org:masterfrom
Silver267:master

Commits on May 10, 2025

Initial commit

Merge branch 'ggml-org:master' into master

Commits on May 11, 2025

Decouple smoothing from temp_ext

Revert changes to the implementation of temp_ext

Merge branch 'ggml-org:master' into master

Added docs

Webui stuff

Merge branch 'ggml-org:master' into master

Commits on May 13, 2025

Merge branch 'ggml-org:master' into master

Commits on May 14, 2025

vulkan: workaround FA compile failures on macos (#13517)

scripts : fix compare-llama-bench.py show parameter (#13514)

docs: Update link to ggml-org in multimodal.md (#13513)

webui: Allow pasting file from clipboard (#13526)

webui : use fflate for more deterministic gzip compress (#13525)

vulkan: KHR_coopmat flash attention (#13506)

cmake: simplify vulkan shader test logic (#13263)

server : fix cache_tokens bug with no cache_prompt (#13533)

server : passthrough the /models endpoint during loading (#13535)

fix: Move build_inp_pos to the top of the graph section for build_granite (#13538)

CUDA: faster Deepseek FA, add Turing support (#13435)

llama : fix quantize with dl backends (#13539)

CUDA: fix crash on large batch size for quant. MoE (#13537)

fix: crash when calling `llama_state_get_size` on a context without a KV cache (#13542)

editorconfig : fix trailing whitespace from #13542 (#13546)

sampling: Port of Smooth Sampling / Quadratic Sampling support#13441Silver267 wants to merge 24 commits intoggml-org:masterggml-org/llama.cpp:masterfrom Silver267:masterSilver267/llama.cpp:masterCopy head branch name to clipboard

Commits

Commits on May 10, 2025

Commits on May 11, 2025

Commits on May 13, 2025

Commits on May 14, 2025

sampling: Port of Smooth Sampling / Quadratic Sampling support#13441
Silver267 wants to merge 24 commits intoggml-org:masterfrom
Silver267:master