Skip to content

Actions: ngxson/llama.cpp

Actions

Server

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
server.yml
2,129 workflow runs
2,129 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

ggml-zdnn: fix #15414, activate FP16 and BF16 acceleration and incorr… Server #2236: Commit 40be511 pushed by ngxson
14m 8s master
master
14m 8s
Add docker protocol support for llama-server model loading (#15790) Server #2235: Commit 4bf5549 pushed by ngxson
14m 6s master
master
14m 6s
context : remove redundant explicit casting to the same type (#15948) Server #2234: Commit f4e664f pushed by ngxson
18m 21s master
master
18m 21s
server : adjust prompt similarity thold + add logs (#15913) Server #2233: Commit f088b6a pushed by ngxson
20m 25s master
master
20m 25s
Vulkan iGPU device selection overhaul and PCI ID API support (#15947) Server #2232: Commit 304ac56 pushed by ngxson
20m 38s master
master
20m 38s
vulkan: Make device memory check more portable (#15939) Server #2231: Commit 6c88ad8 pushed by ngxson
14m 21s master
master
14m 21s
Revert "sycl: add usage of enqueue_functions extension (#14244)" (#1… Server #2230: Commit 704d90c pushed by ngxson
12m 46s master
master
12m 46s
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
ProTip! You can narrow down the results and go further in time using created:<2025-09-09 or the other filters available.