upstream-archive/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-21 13:37:29 -05:00
Code Issues Packages Projects Releases Wiki Activity
11,093 Commits 792 Branches 7,521 Tags

Default Branch

master

fb34fc262c · metal : fix mask bounds in flash attention block pre-pass (#29220) · Updated 2026-09-21 12:31:56 -05:00
 ZIP  TAR.GZ

Branches

gg/tests-fix-save-load-cpu

b9c61e054d · test-save-load-state : compare logits with NMSE and feed expected tokens · Updated 2026-09-21 11:37:03 -05:00
7
1
 ZIP  TAR.GZ
allozaur/feat/external-llama-server-backends

81ba5ada2f · ui : list the server tools at startup · Updated 2026-09-21 09:28:43 -05:00
93
95
 ZIP  TAR.GZ
gg/ggml-alloc-refactor-meta

becf793ed6 · ggml : address review comments on alloc_buffer_n · Updated 2026-09-21 07:50:21 -05:00
7
7
 ZIP  TAR.GZ
aman/shared-exp-fusion

0a743e4e5a · CUDA: fuse shared experts into MMVQ · Updated 2026-09-20 13:26:19 -05:00
31
1
 ZIP  TAR.GZ
cisc/cuda-fuse-grouped-experts

bc4b088b2e · hip fix · Updated 2026-09-20 10:10:20 -05:00
30
3
 ZIP  TAR.GZ
allozaur/fix/ui-horizontal-overflow

e00f71208d · ui: wrap long markdown tokens instead of scrolling every block · Updated 2026-09-19 02:28:40 -05:00
51
9
 ZIP  TAR.GZ
0cc4m/vulkan-coopmat-int8

df9bcc16a7 · avoid compiling f16 acc shader variants · Updated 2026-09-18 08:25:56 -05:00
54
3
 ZIP  TAR.GZ
0cc4m/fp8-scaled-mm

c902192589 · add optional activation scales to mul_mat and mul_mat_id tensors · Updated 2026-09-18 07:30:39 -05:00
281
11
 ZIP  TAR.GZ
0cc4m/vulkan-mulmat-bf16-activations

1f65192798 · clean up mul_mat b-type selection · Updated 2026-09-18 02:43:02 -05:00
118
4
 ZIP  TAR.GZ
gg/parallel-test

d093cf703e · wip · Updated 2026-09-18 01:36:15 -05:00
87
1
 ZIP  TAR.GZ
pwilkin/test-cleanup-sample

5f908c4db4 · CI/tests: log declutter for test-chat · Updated 2026-09-17 12:42:13 -05:00
65
1
 ZIP  TAR.GZ
detestify

10532516c3 · CI/tests: refactor argument handling, add --errors-only to emit only WARN/ERR messages · Updated 2026-09-17 11:47:59 -05:00
86
1
 ZIP  TAR.GZ
osimons/ocp_fp8_cuda

4bcf38c610 · Optimize GEMV kernel for FP8 · Updated 2026-09-17 09:16:21 -05:00
281
7
 ZIP  TAR.GZ
allozaur/ui-reuse-markers

b0ed4fa3b8 · ui : Improvements · Updated 2026-09-16 12:18:44 -05:00
93
19
 ZIP  TAR.GZ
allozaur/ui/models-selector

21aecafc8c · ui : sort reasoning panel attributes · Updated 2026-09-16 12:18:32 -05:00
93
17
 ZIP  TAR.GZ
allozaur/ui/discover-dialog

375a5b08e9 · ui : models discover dialog · Updated 2026-09-16 12:07:37 -05:00
93
13
 ZIP  TAR.GZ
allozaur/ui/shared-model-display

27a11a0f6b · ui : render shared model row hints as native titles · Updated 2026-09-16 12:07:15 -05:00
93
12
 ZIP  TAR.GZ
gg/meta-forward-graph-optimize

8fc3cee2ac · meta : forward graph_optimize to underlying backends · Updated 2026-09-16 11:15:53 -05:00
87
1
 ZIP  TAR.GZ
allozaur/ui/download-pipeline

0172e83237 · ui : model download pipeline · Updated 2026-09-16 06:53:34 -05:00
93
8
 ZIP  TAR.GZ
allozaur/ui/markdown-allow-html

6a4dc7da8e · ui : optional sanitized raw HTML in markdown · Updated 2026-09-16 06:53:34 -05:00
93
9
 ZIP  TAR.GZ