Skip to content

Actions: ngxson/llama.cpp

Actions

Server

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
server.yml
1,796 workflow run results
1,796 workflow run results

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

ggml-backend : add GGML_BACKEND_DEVICE_TYPE_IGPU device type (#15797) Server #2229: Commit 360d653 pushed by ngxson
14m 22s master
master
14m 22s
CUDA: larger SRAM reads for tile FA, AMD FP16 dot (#15927) Server #2228: Commit 0e6ff00 pushed by ngxson
13m 22s master
master
13m 22s
nitpick : correct MB to MiB (#15934) Server #2227: Commit df082f5 pushed by ngxson
13m 29s master
master
13m 29s
ggml-cpu : add check for ARM MATMUL_INT8/i8mm support (#15922) Server #2226: Commit 24a6734 pushed by ngxson
15m 16s master
master
15m 16s
kleidiai: fix GGML_ASSERT(*cur_backend_id != -1) failed (#15614) Server #2225: Commit 2b3efea pushed by ngxson
12m 28s master
master
12m 28s
CANN: Disable acl_graph for prefill stage (#15933) Server #2224: Commit c0389db pushed by ngxson
13m 30s master
master
13m 30s
CUDA: Add fastdiv to k_bin_bcast*, giving 1-3% E2E performance (#… Server #2223: Commit 00681df pushed by ngxson
15m 23s master
master
15m 23s
llama : support T5 models with unequal number of encoder-decoder laye… Server #2222: Commit 4f65885 pushed by ngxson
13m 6s master
master
13m 6s
graph : support non-contiguous Q in build_attn_mha (#15908) Server #2221: Commit 6ab397e pushed by ngxson
15m 13s master
master
15m 13s
ggml-cpu : fix padding in ggml_timestep_embedding (#15917) Server #2220: Commit 9de447d pushed by ngxson
12m 53s master
master
12m 53s
metal : make the backend async (#15906) Server #2219: Commit 0f0a3c2 pushed by ngxson
14m 54s master
master
14m 54s
tests : filter out no-ops from coverage report (#15900) Server #2218: Commit e7b6d83 pushed by ngxson
16m 55s master
master
16m 55s
CANN: Add ROPE sin/cos cache for reuse (#15912) Server #2217: Commit 10d8b2b pushed by ngxson
13m 59s master
master
13m 59s
CANN: implement LRU cache for ACL graphs (#15814) Server #2216: Commit 28b5f19 pushed by ngxson
18m 12s master
master
18m 12s
llama : check returned fn ptrs from ggml_backend_reg_get_proc_address… Server #2215: Commit 86587da pushed by ngxson
13m 51s master
master
13m 51s
vulkan: throw the oom error instead of no memory type found (#15905) Server #2214: Commit ae355f6 pushed by ngxson
14m 30s master
master
14m 30s
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