Skip to content

Actions: ngxson/llama.cpp

Actions

Release

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
release.yml
968 workflow runs
968 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

kv-cache : simplify the interface (#13660) Release #93: Commit 797f2ac pushed by ngxson
31m 9s master
master
31m 9s
model : disable SWA for Phi models (#13676) Release #92: Commit b44890d pushed by ngxson
24m 48s master
master
24m 48s
musa: Upgrade MUSA SDK version to rc4.0.1 and use mudnn::Unary::IDENT… Release #91: Commit 3398305 pushed by ngxson
24m 25s master
master
24m 25s
vulkan: fix warnings (#13626) Release #90: Commit fb1cab2 pushed by ngxson
21m 10s master
master
21m 10s
mtmd-helper : bug fix to token batching in mtmd (#13650) Release #89: Commit b7a1746 pushed by ngxson
23m 47s master
master
23m 47s
model : fix llama4 graph (#13663) Release #88: Commit be02396 pushed by ngxson
18m 24s master
master
18m 24s
llama : remove llama_kv_cache_view API + remove deprecated (#13653) Release #87: Commit a4090d1 pushed by ngxson
35m 11s master
master
35m 11s
CUDA: skip fully masked-out KV in FA vec kernel (#13584) Release #86: Commit b69f164 pushed by ngxson
18m 57s master
master
18m 57s
tests : avoid github urls due to throttling (#13654) Release #85: Commit 759e37b pushed by ngxson
16m 36s master
master
16m 36s
sycl: disable reorder for sycl mulmat (#13536) Release #84: Commit 4245e62 pushed by ngxson
19m 28s master
master
19m 28s
Set GLM4 blk.*.attn_output.weight, kqv_out-* matmul to GGML_PREC_F32 … Release #83: Commit c9c64de pushed by ngxson
32m 18s master
master
32m 18s
metal : fix typo in FA kernel comments (#13651) Release #82: Commit c00a263 pushed by ngxson
20m 39s master
master
20m 39s
kv-cache : add SWA support (#13194) Release #81: Commit e298d2f pushed by ngxson
32m 20s master
master
32m 20s
sycl : Overcoming workaround for mmap() allocation on Windows (#13482) Release #80: Commit f7c9429 pushed by ngxson
29m 43s master
master
29m 43s
common : add load_progress_callback (#13617) Release #79: Commit 1dfbf2c pushed by ngxson
39m 32s master
master
39m 32s
Vulkan: Add f32 accumulator support to quantized mul mat to fix GLM4 … Release #78: Commit 8960efd pushed by ngxson
19m 24s master
master
19m 24s
mtmd : add vision support for llama 4 (#13282) Release #77: Commit 92ecdcc pushed by ngxson
29m 30s master
master
29m 30s
ci : upgraded oneAPI version in SYCL workflows and dockerfile (#13532) Release #76: Commit f71f40a pushed by ngxson
24m 44s master
master
24m 44s
sync : ggml Release #75: Commit d30cb5a pushed by ngxson
29m 0s master
master
29m 0s
CANN: Support MOE Model MUL_MAT_ID (#13042) Release #74: Commit 33d7aed pushed by ngxson
50m 22s master
master
50m 22s
server : added --no-prefill-assistant flag (#13608) Release #73: Commit 6a2bc8b pushed by ngxson
29m 7s master
master
29m 7s
cmake: use the current build config for vulkan-shaders-gen (#13595) Release #72: Commit e3a7cf6 pushed by ngxson
31m 13s master
master
31m 13s
vulkan: move common FA code to flash_attn_base.comp (#13556) Release #71: Commit 2f5a4e1 pushed by ngxson
21m 35s master
master
21m 35s
vulkan: use scalar FA rather than coopmat2 when N==1 (#13554) Release #70: Commit 4f41ee1 pushed by ngxson
17m 46s master
master
17m 46s
llguidance : official v0.7.20 release (no actual changes) [noci] (#13… Release #69: Commit 3e0be1c pushed by ngxson
36m 38s master
master
36m 38s
ProTip! You can narrow down the results and go further in time using created:<2025-05-16 or the other filters available.