Actions: ggml-org/llama.cpp
Actions
14,591 workflow runs
14,591 workflow runs
kv-cache: Fix state restore fragmented cache Server #23899: Pull request #17982 opened by ssweens
webui: fix chat header width when sidebar is closed Server #23898: Pull request #17981 synchronize by polydecay
webui: fix chat header width when sidebar is closed Server #23897: Pull request #17981 synchronize by polydecay
webui: fix chat header width when sidebar is closed Server #23896: Pull request #17981 opened by polydecay
mtmd: refactor audio preprocessing Server #23894: Pull request #17978 opened by ngxson
ggml-hexagon: Implement true Q8_0 quantization on Hexagon NPU for more accurate mixed-precision matmul operations Server #23893: Pull request #17977 opened by ngdxzy
ggml-hexagon: gelu operation Server #23892: Pull request #17921 synchronize by joeldushouyu
Webui: Disable attachment button and model selector button when prompt textbox is disabled. Server #23891: Pull request #17925 synchronize by dariusjlukas
Webui: Disable attachment button and model selector button when prompt textbox is disabled. Server #23890: Pull request #17925 synchronize by dariusjlukas
common : add llama-completion to completion-bash executables Server #23889: Pull request #17976 opened by CISC
common : skip model validation when --completion-bash is requested Server #23888: Pull request #17975 opened by CISC
llama_context: synchronize before reallocating output buffer Server #23887: Pull request #17974 opened by jeffbolznv
clip: move model cgraphs into their own files (#17965) Server #23886: Commit e39a2ce pushed by ngxson
clip: move model cgraphs into their own files Server #23885: Pull request #17965 synchronize by ngxson
Webui: Disable attachment button and model selector button when prompt textbox is disabled. Server #23884: Pull request #17925 synchronize by dariusjlukas
cmake: correct scope - link ws2_32 for MinGW/w64devkit builds in cpp-httplib Server #23883: Pull request #17972 opened by gustrd
Webui: Disable attachment button and model selector button when prompt textbox is disabled. Server #23882: Pull request #17925 synchronize by dariusjlukas
Build CUDA architectures 120 and 121 by default (RTX5000 and GB10) Server #23881: Pull request #17970 opened by DaAwesomeP
llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization Server #23880: Pull request #16653 synchronize by JohannesGaessler
common: support negated args Server #23879: Pull request #17919 synchronize by ngxson
webui: Improve copy to clipboard with text attachments Server #23878: Pull request #17969 synchronize by allozaur
webui: Improve copy to clipboard with text attachments Server #23877: Pull request #17969 opened by allozaur
ggml-hexagon: gelu operation Server #23876: Pull request #17921 synchronize by joeldushouyu
common: support negated args Server #23875: Pull request #17919 synchronize by ngxson