Default Branch

41abbfd599 · qwen4exp: enable rms_norm + mul fusion (#28896) · Updated 2026-09-14 16:16:30 +02:00

Branches

04a134c70b · ci : make release workflows use a deply key · Updated 2026-08-17 08:58:11 +02:00

509
1

051139f511 · conversion: refactor moe exp tensor handling · Updated 2026-08-17 01:54:08 +02:00

514
1

d31babce1e · ui: npm run format for new upstream formatting rules · Updated 2026-08-16 09:08:43 +02:00

518
3

27c299a6ec · wip · Updated 2026-08-15 22:40:41 +02:00

520
1

9871d51fd7 · swap worker <--> main thread design · Updated 2026-08-15 16:00:56 +02:00

522
2

14ae5ffc7b · messages_inp_normalizer · Updated 2026-08-15 14:47:40 +02:00

525
3

be16bac40b · ggml: (cuda) support ggml_rope_set_offset · Updated 2026-08-15 12:35:06 +02:00

525
5

045ec92f2d · update skill · Updated 2026-08-14 18:18:09 +02:00

537
2

fd6e36310f · revert gguf fix · Updated 2026-08-14 17:47:26 +02:00

537
3

52e9954892 · cont : fix · Updated 2026-08-14 16:15:50 +02:00

538
8

7c26c91500 · set coarser granularity for head splits · Updated 2026-08-14 15:10:38 +02:00

538
2

bcc7c630f8 · metal: mul_mat: fix src1 row stride edge cases with has_tensor · Updated 2026-08-14 14:28:08 +02:00

540
1

3d6761b8f5 · rm test · Updated 2026-08-14 12:03:50 +02:00

551
6

2ab78db355 · wip · Updated 2026-08-13 15:20:01 +02:00

580
5

b7d9c4ccaf · nits · Updated 2026-08-12 17:19:57 +02:00

610
20

f370471abf · adapt to upstream changes · Updated 2026-08-11 14:00:58 +02:00

601
2

f236e272f0 · rm gen_flow_temp · Updated 2026-08-11 01:01:16 +02:00

613
18

35d6100372 · ci: add pr-draft-label · Updated 2026-08-09 13:50:12 +02:00

635
1

d483b9b3aa · x-tool-docker --> x-tool-runtime · Updated 2026-08-08 01:48:51 +02:00

642
11

ae86645c28 · Consistify polling frequency of llama-bench with args · Updated 2026-08-07 20:59:38 +02:00

643
1