Skip to content

Actions: ngxson/llama.cpp

Actions

Release

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
release.yml
1,074 workflow runs
1,074 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

ggml : Q2k interleaving implementation - x86/x64 SIMD (#14373) Release #449: Commit baad948 pushed by ngxson
14m 15s master
master
14m 15s
graph : fix equal_seq() check (#14986) Release #448: Commit ba42794 pushed by ngxson
13m 29s master
master
13m 29s
docker : add cann build pipline (#14591) Release #447: Commit 2860d47 pushed by ngxson
14m 43s master
master
14m 43s
quantize : skip tensor override when in fallback mode (#14995) Release #446: Commit daf2dd7 pushed by ngxson
14m 30s master
master
14m 30s
llama : add simple option to enable CPU for MoE weights (--cpu-moe) (… Release #445: Commit a06ed5f pushed by ngxson
26m 58s master
master
26m 58s
Fix params bug in diffusion example (#14993) Release #444: Commit 7845240 pushed by ngxson
17m 45s master
master
17m 45s
llama : allow other bufts when overriding to CPU, add --no-repack opt… Release #443: Commit d6818d0 pushed by ngxson
17m 46s master
master
17m 46s
Vulkan: Fix minor debug mode issues (#14899) Release #442: Commit e08a988 pushed by ngxson
17m 20s master
master
17m 20s
mtmd : support MiniCPM-V 4.0 (#14983) Release #441: Commit 952a47f pushed by ngxson
13m 10s master
master
13m 10s
server : implement universal assisted decoding (#12635) Release #440: Commit 94933c8 pushed by ngxson
37m 48s master
master
37m 48s
llama : merge build_moe_ffn_from_probs function into build_moe_ffn (#… Release #439: Commit c1dacaa pushed by ngxson
42m 5s master
master
42m 5s
server : add openai-style logit_bias support (#14946) Release #438: Commit a9f77a8 pushed by ngxson
31m 17s master
master
31m 17s
Add LLaDA 8b Diffusion model (#14771) Release #437: Commit 8a4a856 pushed by ngxson
38m 3s master
master
38m 3s
graph : reduce splits for recurrent and hybrid models (#14825) Release #436: Commit 66625a5 pushed by ngxson
12m 25s master
master
12m 25s
opencl: add mul_mat_f32_f32_l4_lm and mul_mat_f16_f32_l4_lm (#14809) Release #435: Commit 6e67254 pushed by ngxson
19m 26s master
master
19m 26s
quantize : fix using combined imatrix GGUFs (multiple datasets) (#14973) Release #434: Commit e9192be pushed by ngxson
18m 9s master
master
18m 9s
server : add support for embd_normalize parameter (#14964) Release #433: Commit 41e78c5 pushed by ngxson
1h 13m 27s master
master
1h 13m 27s
HIP: enable mfma mmq on gfx908 and gfx90a for select datatypes and sh… Release #432: Commit ad4a700 pushed by ngxson
59m 11s master
master
59m 11s
sync : ggml Release #431: Commit e32a4ec pushed by ngxson
1h 24m 49s master
master
1h 24m 49s
CUDA: skip masked KV slices for all FA kernels (#14924) Release #430: Commit 92b8810 pushed by ngxson
1h 26m 28s master
master
1h 26m 28s
tests : update for LLAMA_SET_ROWS=1 (#14961) Release #429: Commit 00131d6 pushed by ngxson
17m 4s master
master
17m 4s
graph : fix stack-use-after-return (#14960) Release #428: Commit 1e15bfd pushed by ngxson
26m 56s master
master
26m 56s
embeddings: fix extraction of CLS pooling results (#14927) Release #427: Commit a118d80 pushed by ngxson
47m 52s master
master
47m 52s
HIP: remove the use of __HIP_PLATFORM_AMD__, explicitly support only … Release #426: Commit aa79524 pushed by ngxson
1h 5m 50s master
master
1h 5m 50s
HIP: add GGML_HIP_MMQ_MFMA option to allow disableing the MFMA path. … Release #425: Commit b77d111 pushed by ngxson
54m 26s master
master
54m 26s
ProTip! You can narrow down the results and go further in time using created:<2025-07-29 or the other filters available.