anodyine pushed to codex-dlo-rank-local-gif at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
b1f8847f56 docs: add DLO DP behavior animation
anodyine created branch codex-dlo-lifecycle-optimization in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-dlo-lifecycle-optimization at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
1843f05ced [Model] Optimize MiniMax-H3 DLO component lifecycle
anodyine created branch codex-dlo-rank-local-gif in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-consolidate-lingbot-text-to-video at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
f82ec476bf [Misc] Address LingBot runner review feedback
0a593bdded [Misc] Consolidate LingBot text-to-video runner
Compare 2 commits »
anodyine pushed to codex-design-docs-reorg at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
ac67e8874d Update docs/design/index.md
dfe6e13cfc Update docs/.nav.yml
9d80efcbe5 docs: keep design reorg arrangement-only
ddbbfd29e5 docs: reorganize design navigation
Compare 4 commits »
anodyine created branch codex-diffusion-shutdown-escalation in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine created branch codex-dlo-cache-budget-config in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-dlo-cache-budget-config at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
e1c4b6153a [Model] Make the dummy-run recipe feature-neutral
33297bdb9d [Model] Fold the DLO warmup recipe into _make_dummy_request
f4394f1d3d [Model] Prime DLO cache retention with a startup warmup
c9bb65902e [Model] Learn the retention budget directly from observed peaks
b78db6036a [Model] Fix adaptation test fixture so the capacity cap does not bind
Compare 8 commits »
anodyine created branch codex-dlo-design-docs in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-dlo-design-docs at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
a5ee63fb1f docs: restore quantization overview page
3335e76c97 docs: add quantization design document
9167599b3c docs: rename diffusion CPU offload page
99f946210b docs: rename quantization overview page
ec4e12f379 docs: remove Ray-based execution page
Compare 6 commits »
anodyine created branch codex-dlo-encoder-residency in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-dlo-encoder-residency at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
7db3a0cd5e [Performance] Add per-component DLO offload policy
anodyine created branch codex-dlo-fp8-cache-phase1 in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-dlo-fp8-cache-phase1 at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
1ae56cfcfb Add normalized online FP8 cache for DLO
anodyine created branch codex-dlo-host-weight-plan in anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
anodyine pushed to codex-dlo-host-weight-plan at anodyine/vllm-omni 2026-10-07 20:48:19 -04:00
65c86dec44 [Docs] Clarify TP1 DLO mmap support scope
5ce549bc57 Merge branch 'main' into codex-dlo-host-weight-plan
Compare 2 commits »
anodyine pushed tag v0.25.0rc1 to anodyine/vllm-omni 2026-10-07 20:48:18 -04:00
anodyine pushed tag v0.26.0 to anodyine/vllm-omni 2026-10-07 20:48:18 -04:00
anodyine pushed to bugfix/qwen3-tts-remove-default-seed at anodyine/vllm-omni 2026-10-07 20:48:18 -04:00
d61ad3ffac fix: remove default Qwen3-TTS seeds