Skip to content

Actions: ngxson/llama.cpp

Actions

CI

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
build.yml
1,050 workflow run results
1,050 workflow run results

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

CUDA: faster Deepseek FA, add Turing support (#13435) CI #1410: Commit 6da34fa pushed by ngxson
54m 42s master
master
54m 42s
fix: Move build_inp_pos to the top of the graph section for build_gra… CI #1409: Commit 5e7d95e pushed by ngxson
1h 3m 48s master
master
1h 3m 48s
server : passthrough the /models endpoint during loading (#13535) CI #1408: Commit 0531744 pushed by ngxson
1h 2m 15s master
master
1h 2m 15s
server : fix cache_tokens bug with no cache_prompt (#13533) CI #1407: Commit 360a9c9 pushed by ngxson
58m 3s master
master
58m 3s
cmake: simplify vulkan shader test logic (#13263) CI #1406: Commit 09d13d9 pushed by ngxson
48m 17s master
master
48m 17s
vulkan: KHR_coopmat flash attention (#13506) CI #1405: Commit 24e86ca pushed by ngxson
51m 48s master
master
51m 48s
vulkan: workaround FA compile failures on macos (#13517) CI #1404: Commit ab3971f pushed by ngxson
52m 56s master
master
52m 56s
quantize : improve tensor-type pattern matching (#13033) CI #1403: Commit e5c834f pushed by ngxson
54m 2s master
master
54m 2s
clip : clip.h become private API (⚠️ breaking change) (#13510) CI #1402: Commit 71bdbdb pushed by ngxson
1h 3m 40s master
master
1h 3m 40s
metal : use FA-vec kernel up to batch size 20 (#13496) CI #1401: Commit f0995d2 pushed by ngxson
1h 2m 15s master
master
1h 2m 15s
mtmd : remove libllava, remove clip-quantize-cli (⚠️ breaking change)… CI #1400: Commit b472634 pushed by ngxson
56m 22s master
master
56m 22s
model : Granite MoE shared (#13269) CI #1399: Commit d590cd4 pushed by ngxson
1h 2m 28s master
master
1h 2m 28s
llama-bench : add defrag-thold, check for invalid ranges (#13487) CI #1398: Commit cf0a43b pushed by ngxson
53m 54s master
master
53m 54s
opencl: remove unnecessary assert for add (#13257) CI #1397: Commit f0d46ef pushed by ngxson
54m 46s master
master
54m 46s
clip : cap max image size 1024 for qwen vl model (#13478) CI #1396: Commit de4c07f pushed by ngxson
1h 14m 40s master
master
1h 14m 40s
llama/ggml: add LLM training support (#10544) CI #1395: Commit 10d2af0 pushed by ngxson
1h 8m 47s master
master
1h 8m 47s
context : fix state io for memory-less contexts (#13470) CI #1394: Commit 064cc59 pushed by ngxson
1h 2m 51s master
master
1h 2m 51s
server : allow content to be null in oaicompat_completion_params_pars… CI #1393: Commit 91159ee pushed by ngxson
51m 13s master
master
51m 13s
llama-bench : accept ranges for integer parameters (#13410) CI #1392: Commit 22cdab3 pushed by ngxson
55m 58s master
master
55m 58s
CUDA: fix misaligned synchronization in FA (#13469) CI #1391: Commit 95e1888 pushed by ngxson
57m 15s master
master
57m 15s
ggml : add mrope kernel for metal (#13457) CI #1390: Commit df84919 pushed by ngxson
56m 27s master
master
56m 27s
enable dpcpp nightly builds with libraries (#13406) CI #1389: Commit 1449214 pushed by ngxson
58m 55s master
master
58m 55s
mtmd : Use RMS norm for InternVL 3 38B and 78B mmproj (#13459) CI #1388: Commit c104023 pushed by ngxson
56m 47s master
master
56m 47s
tools : fix uninitialized llama_batch in server (#13436) CI #1387: Commit 9a390c4 pushed by ngxson
51m 24s master
master
51m 24s
CUDA: fix crash with partial offloading of MoE (#13439) CI #1386: Commit 7474e00 pushed by ngxson
51m 33s master
master
51m 33s