Default Branch

f818065d75 · CUDA: batch out_prod broadcast (dps2>1) path with cublasSgemmBatched (#24426) · Updated 2026-06-26 05:51:25 +00:00

Branches

946ede7372 · add windows-openvino to check-release · Updated 2026-06-25 22:24:26 +00:00

3
1

81313a35ae · type check for get_arr_int · Updated 2026-06-25 16:54:57 +00:00

8
4

2e4cbade70 · Merge branch 'master' into xsn/mtmd_ds_ocr_tiles · Updated 2026-06-25 14:28:50 +00:00

8
10

68ed5149fb · bring back examples, add mtmd · Updated 2026-06-25 13:23:03 +00:00

42
2

bf05250df9 · use unsigned ints · Updated 2026-06-25 13:02:50 +00:00

11
3

3199d5357c · chat: harden caps check · Updated 2026-06-24 13:16:42 +00:00

25
1

a14f8d2ed5 · fix test case · Updated 2026-06-24 11:38:25 +00:00

25
3

ef687feb42 · common: remove unused json-partial · Updated 2026-06-24 10:49:42 +00:00

26
1

a432e6f863 · use destructor instead · Updated 2026-06-23 20:57:20 +00:00

35
10

095058ca19 · add arg --threads-sampling · Updated 2026-06-22 18:03:49 +00:00

44
4

1b82e9ae51 · fix windows · Updated 2026-06-22 14:20:56 +00:00

46
7

037397792a · vulkan: split ggml-vulkan.cpp file · Updated 2026-06-22 13:50:01 +00:00

47
1

bec3083830 · metal : per-op source split + parallel compile (#24021) · Updated 2026-06-22 11:15:48 +00:00

47
1

7ac864bf97 · disable DEBUG_TIMINGS · Updated 2026-06-21 11:38:09 +00:00

65
15

f1ef61fb1b · server: add "verbose" field to schema · Updated 2026-06-21 09:16:06 +00:00

59
1

447b0c3646 · poc: threadpool sampling · Updated 2026-06-20 20:08:42 +00:00

65
14

5a7462237e · remove duplicated init calls · Updated 2026-06-19 09:07:38 +00:00

83
18

37db4fa4be · improve test · Updated 2026-06-17 15:42:56 +00:00

122
2

42874dfd8f · clean up logging and timing · Updated 2026-06-17 11:47:53 +00:00

193
6

fcff47bcb1 · Merge branch 'master' into add-long-debug-prompt · Updated 2026-06-15 15:05:22 +00:00

149
2