Skip to content

Actions: ngxson/llama.cpp

Actions

Server

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
server.yml
2,131 workflow runs
2,131 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

vulkan: Fix OOB accesses in soft_max_back (#15861) Server #2213: Commit 4f63cd7 pushed by ngxson
26m 8s master
master
26m 8s
HIP: use v_dot2_f32_f16 instruction for FA (#15884) Server #2212: Commit 17bc5a8 pushed by ngxson
24m 6s master
master
24m 6s
Workaround for subgroup arithmetic failing on MoltenVK with AMD GPUs … Server #2211: Commit ed54e32 pushed by ngxson
12m 51s master
master
12m 51s
CUDA: Add mul_mat_id support for the mmf kernel (#15767) Server #2210: Commit a972fae pushed by ngxson
14m 6s master
master
14m 6s
CUDA: fix GET_ROWS for large tensors (#15882) Server #2209: Commit 550cf72 pushed by ngxson
15m 53s master
master
15m 53s
requirements : update transformers/torch for Embedding Gemma (#15828) Server #2208: Commit 70cd37d pushed by ngxson
13m 0s master
master
13m 0s
json : support enum values within allOf (#15830) Server #2207: Commit 7057faf pushed by ngxson
20m 1s master
master
20m 1s
vulkan: sort graph to allow more parallel execution (#15850) Server #2206: Commit e68aa10 pushed by ngxson
16m 9s master
master
16m 9s
chat : Deepseek V3.1 reasoning and tool calling support (OpenAI Style… Server #2205: Commit 8802156 pushed by ngxson
26m 5s master
master
26m 5s
server : bring back timings_per_token (#15879) Server #2204: Commit 56920f5 pushed by ngxson
13m 52s master
master
13m 52s
metal : refactor + optimize (#15857) Server #2203: Commit f28d4f4 pushed by ngxson
12m 37s master
master
12m 37s
CUDA: non-contiguous src0 not supported for PAD (#15869) Server #2202: Commit 5ef22d2 pushed by ngxson
22m 8s master
master
22m 8s
batched-bench : fix llama_synchronize usage during prompt processing … Server #2201: Commit a885dcf pushed by ngxson
13m 22s master
master
13m 22s
tests: large sizes for get_rows (#15687) Server #2200: Commit d413dca pushed by ngxson
18m 18s master
master
18m 18s
CANN: Stream sync between devices for acl_graph (#15809) Server #2199: Commit 85ca66a pushed by ngxson
13m 22s master
master
13m 22s
vulkan: support im2col_3d (#15795) Server #2198: Commit 3976dfb pushed by ngxson
14m 15s master
master
14m 15s
ggml-cpu: clean up s390x SIMD (#15855) Server #2197: Commit d36e61c pushed by ngxson
13m 23s master
master
13m 23s
vulkan: Support pad_ext (#15794) Server #2196: Commit c97b5e5 pushed by ngxson
23m 21s master
master
23m 21s
vulkan: Use larger loads in scalar/coopmat1 matmul (#15729) Server #2195: Commit 267e998 pushed by ngxson
21m 34s master
master
21m 34s
ggml WebGPU: remove userdata from request adapter callback (#15527) Server #2194: Commit 3b15924 pushed by ngxson
19m 2s master
master
19m 2s
CUDA: faster tile FA (Pascal/AMD), headsize 256 (#15769) Server #2193: Commit 79bc429 pushed by ngxson
14m 1s master
master
14m 1s
kleidiai: generalize compute_forward_kv_cache to compute_forward_fp16… Server #2192: Commit c4df49a pushed by ngxson
13m 15s master
master
13m 15s
server : speed up tests (#15836) Server #2191: Commit 3c3635d pushed by ngxson
13m 37s master
master
13m 37s
server : implement prompt processing progress report in stream mode (… Server #2190: Commit 61bdfd5 pushed by ngxson
12m 54s master
master
12m 54s
ggml-cpu: drop support for nnpa intrinsics (#15821) Server #2189: Commit 186415d pushed by ngxson
12m 44s master
master
12m 44s
ProTip! You can narrow down the results and go further in time using created:<2025-09-06 or the other filters available.