Skip to content

Actions: Dillflix/llama.cpp

Actions

Server

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
server.yml
10 workflow runs
10 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
CUDA: faster tile FA (Pascal/AMD), headsize 256 (#15769) Server #10: Commit 79bc429 pushed by Dillflix
13m 22s master
master
13m 22s
tests : fix test-opt with GGML_BACKEND_DL (#15599) Server #9: Commit bcbddcd pushed by Dillflix
25m 16s master
master
25m 16s
vulkan: apply MUL_MAT_ID subgroup optimization to non-coopmat devices… Server #8: Commit 043fb27 pushed by Dillflix
13m 43s master
master
13m 43s
test-opt: allow slight inprecision (#15503) Server #7: Commit e92734d pushed by Dillflix
13m 42s master
master
13m 42s
sched : fix possible use of wrong ids tensor when offloading moe prom… Server #6: Commit 54a241f pushed by Dillflix
13m 3s master
master
13m 3s
CUDA: replace GGML_CUDA_F16 with CUDA arch checks (#15433) Server #5: Commit 7a6e91a pushed by Dillflix
12m 9s master
master
12m 9s
common : fix double bos, use common_chat_templates for add_bos and ad… Server #4: Commit 5e6229a pushed by Dillflix
21m 36s master
master
21m 36s
CUDA: fix negative KV_max values in FA (#15321) Server #3: Commit 4227c9b pushed by Dillflix
13m 3s master
master
13m 3s
llama : add 18-layer model type for Gemma 3-270m (#15319) Server #2: Commit 7a0de96 pushed by Dillflix
23m 2s master
master
23m 2s
cuda: refactored ssm_scan and use CUB (#13291) Server #1: Commit 79c1160 pushed by Dillflix
15m 0s master
master
15m 0s