Skip to content

Actions: ngxson/llama.cpp

Actions

CI

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
build.yml
2,218 workflow runs
2,218 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

HIP: use v_dot2_f32_f16 instruction for FA (#15884) CI #2115: Commit 17bc5a8 pushed by ngxson
1h 15m 16s master
master
1h 15m 16s
Workaround for subgroup arithmetic failing on MoltenVK with AMD GPUs … CI #2114: Commit ed54e32 pushed by ngxson
1h 11m 16s master
master
1h 11m 16s
CUDA: Add mul_mat_id support for the mmf kernel (#15767) CI #2113: Commit a972fae pushed by ngxson
1h 11m 12s master
master
1h 11m 12s
CUDA: fix GET_ROWS for large tensors (#15882) CI #2112: Commit 550cf72 pushed by ngxson
1h 11m 34s master
master
1h 11m 34s
requirements : update transformers/torch for Embedding Gemma (#15828) CI #2111: Commit 70cd37d pushed by ngxson
1h 17m 38s master
master
1h 17m 38s
json : support enum values within allOf (#15830) CI #2110: Commit 7057faf pushed by ngxson
1h 5m 18s master
master
1h 5m 18s
vulkan: sort graph to allow more parallel execution (#15850) CI #2109: Commit e68aa10 pushed by ngxson
1h 9m 7s master
master
1h 9m 7s
chat : Deepseek V3.1 reasoning and tool calling support (OpenAI Style… CI #2108: Commit 8802156 pushed by ngxson
1h 16m 43s master
master
1h 16m 43s
server : bring back timings_per_token (#15879) CI #2107: Commit 56920f5 pushed by ngxson
1h 7m 23s master
master
1h 7m 23s
metal : refactor + optimize (#15857) CI #2106: Commit f28d4f4 pushed by ngxson
1h 7m 33s master
master
1h 7m 33s
CUDA: non-contiguous src0 not supported for PAD (#15869) CI #2105: Commit 5ef22d2 pushed by ngxson
1h 5m 26s master
master
1h 5m 26s
batched-bench : fix llama_synchronize usage during prompt processing … CI #2104: Commit a885dcf pushed by ngxson
1h 15m 16s master
master
1h 15m 16s
tests: large sizes for get_rows (#15687) CI #2103: Commit d413dca pushed by ngxson
1h 6m 38s master
master
1h 6m 38s
CANN: Stream sync between devices for acl_graph (#15809) CI #2102: Commit 85ca66a pushed by ngxson
1h 5m 42s master
master
1h 5m 42s
vulkan: support im2col_3d (#15795) CI #2101: Commit 3976dfb pushed by ngxson
1h 11m 29s master
master
1h 11m 29s
ggml-cpu: clean up s390x SIMD (#15855) CI #2100: Commit d36e61c pushed by ngxson
1h 15m 2s master
master
1h 15m 2s
vulkan: Support pad_ext (#15794) CI #2099: Commit c97b5e5 pushed by ngxson
1h 9m 37s master
master
1h 9m 37s
vulkan: Use larger loads in scalar/coopmat1 matmul (#15729) CI #2098: Commit 267e998 pushed by ngxson
1h 2m 59s master
master
1h 2m 59s
ggml WebGPU: remove userdata from request adapter callback (#15527) CI #2097: Commit 3b15924 pushed by ngxson
1h 24m 23s master
master
1h 24m 23s
CUDA: faster tile FA (Pascal/AMD), headsize 256 (#15769) CI #2096: Commit 79bc429 pushed by ngxson
1h 11m 35s master
master
1h 11m 35s
kleidiai: generalize compute_forward_kv_cache to compute_forward_fp16… CI #2095: Commit c4df49a pushed by ngxson
1h 15m 18s master
master
1h 15m 18s
server : implement prompt processing progress report in stream mode (… CI #2094: Commit 61bdfd5 pushed by ngxson
1h 11m 31s master
master
1h 11m 31s
ggml-cpu: drop support for nnpa intrinsics (#15821) CI #2093: Commit 186415d pushed by ngxson
1h 14m 39s master
master
1h 14m 39s
aLoRA Support (#15327) CI #2092: Commit fd62188 pushed by ngxson
1h 14m 0s master
master
1h 14m 0s
Thinking model disabled assistant prefill (#15404) CI #2091: Commit 5fac79c pushed by ngxson
1h 14m 57s master
master
1h 14m 57s
ProTip! You can narrow down the results and go further in time using created:<2025-09-05 or the other filters available.