Skip to content

Actions: srogmann/llama.cpp

Actions

Release

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
release.yml
64 workflow runs
64 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
chat: fix case where template accepts type content only (#19419) Release #39: Commit 98e57ca pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
model : support Step3.5-Flash (#19283) Release #38: Commit b831118 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
vulkan: make FA mask/softcap enables spec constants (#19309) Release #37: Commit f9bd518 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
metal : adaptive CPU/GPU interleave based on number of nodes (#19369) Release #36: Commit 22cae83 pushed by srogmann
1d 0h 0m 3s master
master
1d 0h 0m 3s
sampling : delegate input allocation to the scheduler (#19266) Release #35: Commit faa1bc2 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
vulkan: disable coopmat1 fa on Nvidia Turing (#19290) Release #34: Commit 32b17ab pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
ggml-hexagon: flash-attention and reduce-sum optimizations (#19141) Release #33: Commit 89f10ba pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
lookup, lookahead: fix crash when n_ctx not specified (#18729) Release #32: Commit 1488339 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
hexagon: enable offloading to Hexagon on Windows on Snapdragon (#19150) Release #31: Commit ce38a4d pushed by srogmann
1d 0h 0m 3s master
master
1d 0h 0m 3s
graph : fix nkvo offload with FA (#19105) Release #30: Commit 8f80d1b pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
CUDA: re-use MLA K data for V in MMA FA (#19057) Release #29: Commit 8f91ca5 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
ggml : extend ggml_pool_1d + metal (#16429) Release #28: Commit 388ce82 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
vulkan: Implement mmvq for iq1_s/iq1_m (#18450) Release #27: Commit 706e3f9 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
model: support youtu-vl model (#18479) Release #26: Commit ced765b pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
sync : ggml Release #25: Commit 13814eb pushed by srogmann
1d 0h 0m 3s master
master
1d 0h 0m 3s
Webui/prompt processing progress (#18300) Release #24: Commit c9a3b40 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
android: routine maintenance - Dec 2025 (#18338) Release #23: Commit c136605 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
server : handle closed connection for tasks (#18459) Release #22: Commit 2a85f72 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
HIP: Use mmq on MFMA devices for MUL_MAT_ID in cases where a lot of s… Release #21: Commit 4ffc47c pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
model: add llama 4 scaling for mistral-large (deepseek arch) (#17744) Release #20: Commit 4d37262 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
opencl: add sqr, sqrt, mean and ssm_conv (#17476) Release #19: Commit 7cba58b pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
Revive MUL_MAT_ID to perf testing (#17397) Release #18: Commit 3f3a4fb pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
vulkan: implement ADD1, ARANGE, FILL, SOFTPLUS, STEP, ROUND, CEIL, FL… Release #17: Commit 7d77f07 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
server: split HTTP into its own interface (#17216) Release #16: Commit 0de8878 pushed by srogmann
1d 0h 0m 2s master
master
1d 0h 0m 2s
mtmd : fix embedding size for image input (#17123) Release #15: Commit b8595b1 pushed by srogmann
1d 0h 0m 1s master
master
1d 0h 0m 1s