upstream-archive/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-10-01 18:37:28 -05:00
Code Issues Packages Projects Releases Wiki Activity
11,345 Commits 829 Branches 7,706 Tags

Default Branch

master

a868c3e3c5 · hexagon: add q2_k and q3_k quant type support (#29717) · Updated 2026-10-01 16:12:25 -05:00
 ZIP  TAR.GZ

Branches

pr-24546-recreated

f01a75498e · CUDA: pick MMQ tile size against ncols_opt set on the host side · Updated 2026-09-07 09:39:17 -05:00
496
2
 ZIP  TAR.GZ
revert-mmq-moe-ncols

2265b7fa41 · Revert "CUDA: size routed MoE MMQ N-tiles from typical expert width on RDNA3 (#24546)" · Updated 2026-09-07 09:19:36 -05:00
497
1
 ZIP  TAR.GZ
xsn/models_fix_hang

eee2fa9c79 · server: keep a queued model out of the victim pool until its waiters leave · Updated 2026-09-07 06:39:02 -05:00
508
2
 ZIP  TAR.GZ
aman/draft-single-device

921c3e1aac · llama-context: add warning if sm tensor is used with one device · Updated 2026-09-06 07:21:21 -05:00
537
5
 ZIP  TAR.GZ
aman/cuda-fix-races

f0f9902d4c · use ggml_cuda_syncwarp · Updated 2026-09-06 03:08:55 -05:00
527
2
 ZIP  TAR.GZ
allozaur/ui/chat-perf

1cd349c00d · refactor : address review remarks · Updated 2026-09-05 20:29:20 -05:00
525
15
 ZIP  TAR.GZ
nikwen/m2-max-tunings

5150132b9a · metal : add remaining fa-vec tunings for M2 Max · Updated 2026-09-05 16:16:07 -05:00
525
1
 ZIP  TAR.GZ
xsn/log_jsonl

4c31a6aac2 · rename unknown to none · Updated 2026-09-05 14:52:33 -05:00
526
2
 ZIP  TAR.GZ
allozaur/discover-wip

5efdc8877b · ui : hoist the duplicated sidecar badge classes in the quant chip · Updated 2026-09-04 13:07:38 -05:00
530
84
 ZIP  TAR.GZ
nikwen/metal-leak

b6cc43272d · metal : fix memory leak in early return · Updated 2026-09-04 11:37:43 -05:00
533
1
 ZIP  TAR.GZ
gg/spec-default-by-default

7c2e0dea43 · common : enable default speculative config · Updated 2026-09-04 08:25:14 -05:00
566
1
 ZIP  TAR.GZ
xsn/synthid

19f255903c · add check script · Updated 2026-09-03 18:17:09 -05:00
552
3
 ZIP  TAR.GZ
xsn/mtmd_more_const

3623b0ae48 · mtmd: propagate const to preproc class · Updated 2026-09-03 05:15:22 -05:00
562
1
 ZIP  TAR.GZ
xsn/llama-connect

a2ec6b0fff · server: do not advertise a connect code without the tunnel · Updated 2026-09-03 03:12:11 -05:00
572
8
 ZIP  TAR.GZ
xsn/fix_rebuild_new_commit

ea14703b86 · llama: fix whole source code rebuilt on each new commit · Updated 2026-09-02 18:07:06 -05:00
570
1
 ZIP  TAR.GZ
xsn/idefics_fix

0349b3fbcb · mtmd: fix idefics3 preproc · Updated 2026-09-02 16:56:21 -05:00
572
1
 ZIP  TAR.GZ
xsn/mtmd_tokenize_from_parts

1a257585da · move add_special to call level · Updated 2026-09-02 12:28:24 -05:00
585
3
 ZIP  TAR.GZ
copilot/fix-gpu-cuda-job

7cce890466 · Initial plan · Updated 2026-09-02 11:46:05 -05:00
582
1
 ZIP  TAR.GZ
mmqx-rdna3-routed-moe-tiling

fbc87078d0 · HIP: mmq: add use_typical_moe_ncols to mmq-config-pascal-older · Updated 2026-09-02 10:24:53 -05:00
606
7
 ZIP  TAR.GZ
xsn/dsv4_vision_prob_bias

316ed8aade · nits · Updated 2026-09-02 09:46:27 -05:00
583
2
 ZIP  TAR.GZ