Skip to content

Actions: ngxson/llama.cpp

Actions

CI

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
build.yml
1,082 workflow run results
1,082 workflow run results

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

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
Add --no-op-offload to improve -ot pp perf in MoE models like lla… CI #1385: Commit 7f323a5 pushed by ngxson
51m 55s master
master
51m 55s
mtmd : support InternVL 3 38B and 78B mmproj (#13443) CI #1384: Commit 3eac209 pushed by ngxson
46m 54s master
master
46m 54s
CUDA: fix race conditions FlashAttention kernels (#13438) CI #1383: Commit 0208355 pushed by ngxson
1h 1m 54s master
master
1h 1m 54s
vocab : add ByteDance-Seed/Seed-Coder (#13423) CI #1382: Commit d2a4ef0 pushed by ngxson
57m 39s master
master
57m 39s
mtmd : add hard limit on image resolution for qwen2vl / qwen2.5vl (#1… CI #1381: Commit 15e6125 pushed by ngxson
55m 55s master
master
55m 55s
llguidance : set tokenizer slices to default (#13424) CI #1380: Commit 43dfd74 pushed by ngxson
47m 31s master
master
47m 31s
mtmd : support InternVL 2.5 and 3 (#13422) CI #1379: Commit 053367d pushed by ngxson
58m 28s master
master
58m 28s
CUDA: fix FlashAttention on Turing (#13415) CI #1378: Commit d891942 pushed by ngxson
46m 55s master
master
46m 55s
arg : add env var to control mmproj (#13416) CI #1377: Commit 7fef117 pushed by ngxson
1h 9m 21s master
master
1h 9m 21s
vulkan: scalar flash attention implementation (#13324) CI #1376: Commit dc1d2ad pushed by ngxson
1h 6m 31s master
master
1h 6m 31s
chore(llguidance): use tagged version that does not break the build (… CI #1375: Commit 7c28a74 pushed by ngxson
51m 21s master
master
51m 21s
server : vision support via libmtmd (#12898) CI #1374: Commit 33eff40 pushed by ngxson
43m 55s master
master
43m 55s
sycl : implementation of reordered Q4_0 MMVQ for Intel GPUs (#12858) CI #1373: Commit 17512a9 pushed by ngxson
42m 32s master
master
42m 32s
metal : optimize MoE for large batches (#13388) CI #1372: Commit 611aa91 pushed by ngxson
51m 38s master
master
51m 38s
CUDA: FA support for Deepseek (Ampere or newer) (#13306) CI #1371: Commit 0cf6725 pushed by ngxson
44m 45s master
master
44m 45s
llama : do not crash if there is no CPU backend (#13395) CI #1370: Commit 27ebfca pushed by ngxson
53m 29s master
master
53m 29s
CUDA: fix crash on large batch size for MoE models (#13384) CI #1369: Commit 5c86c9e pushed by ngxson
1h 2m 1s master
master
1h 2m 1s
imatrix : Add --parse-special for enabling parsing of special tokens … CI #1368: Commit efb8b47 pushed by ngxson
44m 26s master
master
44m 26s
llama-run: add support for downloading models from ModelScope (#13370) CI #1367: Commit 0527771 pushed by ngxson
52m 19s master
master
52m 19s
mtmd : fix batch_view for m-rope (#13397) CI #1366: Commit 2189fd3 pushed by ngxson
56m 24s master
master
56m 24s
rpc : add rpc_msg_set_tensor_hash_req (#13353) CI #1365: Commit b486ba0 pushed by ngxson
1h 3m 16s master
master
1h 3m 16s
vulkan: Allow up to 4096 elements for mul_mat_id row_ids (#13326) CI #1364: Commit 02115dc pushed by ngxson
56m 11s master
master
56m 11s