Skip to content

Actions: ngxson/llama.cpp

Actions

Release

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
release.yml
1,141 workflow runs
1,141 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

llama : add support for EmbeddingGemma 300m (#15798) Release #691: Commit fb15d64 pushed by ngxson
26m 34s master
master
26m 34s
metal : Add template specialization for mul_mm_id w/ ne20 == 10 (#15799) Release #690: Commit 856ed09 pushed by ngxson
19m 12s master
master
19m 12s
llama : set n_outputs to 1 to avoid 0 outputs mean-pooling (#15791) Release #689: Commit d1e2adb pushed by ngxson
17m 50s master
master
17m 50s
CANN: Refactor ND to NZ workspace to be per-device (#15763) Release #688: Commit c1c354e pushed by ngxson
20m 14s master
master
20m 14s
server: add exceed_context_size_error type (#15780) Release #687: Commit a68d914 pushed by ngxson
23m 38s master
master
23m 38s
ggml: add ops for WAN video model (cuda && cpu) (#15669) Release #686: Commit 0a1b398 pushed by ngxson
59m 37s master
master
59m 37s
CANN: Fix precision issue on 310I DUO multi-devices (#15784) Release #685: Commit 5421f63 pushed by ngxson
20m 24s master
master
20m 24s
opencl: add hs=40 to FA (#15758) Release #684: Commit 820bc98 pushed by ngxson
22m 34s master
master
22m 34s
CANN: fix acl_rstd allocation size in ggml_cann_rms_norm (#15760) Release #683: Commit 239b60e pushed by ngxson
19m 37s master
master
19m 37s
vulkan: fix mmv subgroup16 selection (#15775) Release #682: Commit dff7551 pushed by ngxson
27m 57s master
master
27m 57s
vulkan: don't use std::string in load_shaders, to improve compile tim… Release #681: Commit 0fce7a1 pushed by ngxson
54m 10s master
master
54m 10s
vulkan : update ggml_vk_instance_validation_ext_available (#15666) Release #680: Commit 8227695 pushed by ngxson
57m 24s master
master
57m 24s
ggml vulkan: add hardsigmoid and hardswish operations (#15762) Release #679: Commit 0014fb4 pushed by ngxson
50m 33s master
master
50m 33s
CUDA: Optimize rms_norm_f32 kernel and its fused variants, giving 1… Release #678: Commit 661ae31 pushed by ngxson
1h 5m 6s master
master
1h 5m 6s
sampling : optimize dist sampler (#15704) Release #677: Commit cdedb70 pushed by ngxson
15m 25s master
master
15m 25s
llama : fix incorrect model type for Gemma 270M (#15764) Release #676: Commit 2c8dac7 pushed by ngxson
23m 13s master
master
23m 13s
CANN: Add RoPE contiguous check for 310I DUP device (#15735) Release #675: Commit 5eae934 pushed by ngxson
19m 19s master
master
19m 19s
ggml-cpu : optimize RVV kernels (#15720) Release #674: Commit 05c0380 pushed by ngxson
19m 37s master
master
19m 37s
CANN: Mask unsupported TRANSPOSE_1D operator (#15733) Release #673: Commit f6da8cb pushed by ngxson
16m 22s master
master
16m 22s
CANN: Fix type float_t to float (#15736) Release #672: Commit 8a2234e pushed by ngxson
57m 4s master
master
57m 4s
fix: resolve unsigned int initialization warning for n_dims/size in g… Release #671: Commit 3de0082 pushed by ngxson
20m 59s master
master
20m 59s
llama: -fa 1/0/-1 aliases for -fa on/off/auto (#15746) Release #670: Commit c466abe pushed by ngxson
24m 51s master
master
24m 51s
vulkan: fix shaders gen when no integer dot is available (#15740) Release #669: Commit 0a2a384 pushed by ngxson
23m 33s master
master
23m 33s
CANN: Resolve soft_max precision issue (#15730) Release #668: Commit 9961d24 pushed by ngxson
16m 38s master
master
16m 38s
vulkan: Fix macro parameter order for f32 matmul shaders (#15716) Release #667: Commit 25f1045 pushed by ngxson
29m 25s master
master
29m 25s
ProTip! You can narrow down the results and go further in time using created:<2025-09-02 or the other filters available.