Commit Graph

Select branches

Hide Pull Requests

ISSUE-2

ISSUE-5

SamitHuang-patch-1

add-skill

agent/daily-pr-labeler-workflow

backup/vllm-align-remote-20260915-0ae9bdd8c

bugfix/qwen3-tts-batched-seeded-mtp

bugfix/qwen3-tts-customvoice-asr-escalation

bugfix/qwen3-tts-remove-default-seed

bugfix/qwen3-tts-seed-propagation

buildkite-2

ci/warmup-full-concurrency-batch

claude/fix-diffusion-cache-ci-iFBqB

codex-6297-topology-aware-diffusion-worker-title-v2

codex-api-server-user-guide

codex-canonical-image-task-examples

codex-community-docs-refresh

codex-consolidate-lingbot-text-to-video

codex-contributor-agent-skills-docs

codex-design-docs-reorg

codex-diffusion-shutdown-escalation

codex-dlo-cache-budget-config

codex-dlo-design-docs

codex-dlo-encoder-residency

codex-dlo-fp8-cache-phase1

codex-dlo-host-weight-plan

codex-dlo-lifecycle-optimization

codex-dlo-rank-local-gif

codex-docs-quantization-recipes

codex-docs-v0.30.0

codex-fix-env-var-ratchet-6217

codex-fix-lingbot-t2i-save

codex-fix-omni-sleep-wake-resume

codex-fix-pr-6011

codex-fix-rtd-doc-link

codex-flux2-klein-bf16-hwr

codex-harden-x-to-video-audio

codex-hwr-build-admission

codex-hwr-capacity-scan

codex-hwr-domain-timeout

codex-hwr-integrity-policy

codex-hwr-memory-diagnostic

codex-hwr-quarantine-cleanup

codex-hwr-registration-policy

codex-issue-7035-h100-determinism

codex-ltx-2-5-diffusers

codex-ltx2-vae-channels-last

codex-minimax-h3-vendor-callback

codex-pd-navigation

codex-pr6279-benchmark-assets

codex-precheck-pr-example-policy

codex-quantization-consistency

codex-quantization-support-guide

codex-qwenimage-vae-spatial-decode-auto

codex-rank-local-mmap

codex-request-vae-fastpaths

codex-revert-mammothmoda2

codex-revert-pr-reviewer

codex-review-config-provenance

codex-split-attention-offload-guides

codex-update-cache-dit-1.5.0

codex-user-guide-feature-alignment

codex-v0.26.0-docs

codex-vae-b300-stack-tiling

codex-vae-compile-warmup

codex-vae-spatial-decode-auto

codex-wan-vae-data-movement

codex/cleanup-pr-template

codex/dlo-registered-mmap-prototype

codex/dlo-runtime-cache-phase-b

codex/fix-diffusion-allrank-rpc

codex/fix-dlo-5954

codex/fix-hub-kernel-dependencies

codex/fix-hunyuan-image3-expert-mapping

codex/fix-mot-one-sided-bias

codex/fix-offline-profiler-config

codex/fix-qwen3-omni-moe-backend

codex/general-example-docs

codex/host-weight-post-load-publication

codex/lingbot-dense-latest-main

codex/module-design-doc-layout

codex/omni-edge-abstraction

codex/qwen3-tts-default-chunk-ramp

codex/readme-v0.24.0-docs

codex/shared-stage-payload-transport

codex/unskip-fp8-z-image-quality

codex/update-architecture-overview

codex/update-codeowners-maintainers

codex/update-readme-slides-link

community-recipes

copilot/discuss-issue-823

copilot/find-opening-prs-excluding-diffusion

copilot/fix-code-review-comment

copilot/fix-output-issues

copilot/fix-typo-in-documentation

copilot/rebase-to-upstream-vllm-v0150

copilot/research-github-issue-4510

copilot/revert-3162-whitelist-optimization-signed

copilot/update-actions-upload-artifact

copilot/update-documentation-in-pull-request-2972

debug/auxiliary-omni-lifecycle

debug/cosyvoice-conditioning

debug/cpu-collective-isolation

debug/engine-worker-ownership

debug/qwen25-two-card-fixture

debug/qwen3-omni-doc-audio-artifacts

debug/qwen3-tts-failed-speech-audio

debug/realtime-invalid-fixtures

debug/request-session-preservation

debug/sana-video-parallel-fixture

debug/stage-client-reload-isolation

dev/migrate-MR-v2

dev/rebase-v0.15.0

dev/rebase-v0.18.0

dev/rebase-v0170

dev/vllm-align

dev/vllm-align-baseline-v0300

dev/voxcpm2-unified-decode-graph

dev/voxcpm2-unified-graph-followups

docker-doc

dockerfile

docs/add-new-committers

docs/repository-owned-review-skill

docs/update-v0.22.0-release

docs/vllm-0.25-bump

feat/diffusion-prefetch-protection

feat/qwen3-tts-deadline-scheduler

feat/qwen3-tts-deadline-scheduler-signed

feat/vllm-upgrade

feature/precheck-pr-skill

feature/refactor-2

fix-mot-rmsnorm-vllm021

fix/cfg-companion-output-race

fix/entrypoint-ci-timeouts

fix/entrypoint-ci-timeouts-main

fix/minicpmo45-duplex-context-overflow-signed

fix/omni-weight-snapshot-prefetch

fix/pr4709-review-followup

fix/qwen3-tts-codec-eos-guard

fix/qwen3-tts-full-payload-buffer

fix/rebase-pipeline-upload-all-e2e

fix/tp-local-kv-recv-consensus

fix/version-tag-alignment-3858

frontend/multi-api-server-scaling

governance

governance-update

h100-ci

helios/accuracy-maintained-main

khluu-patch-1

khluu-patch-2

khluu/ci

khluu/ci_pipeline_gen

khluu/fix_ci

khluu/release

khluu/release-docker-images

khluu/release2

khluu/release3

lance-model-support

lishunyang12/skip-softmax-trtllm-gen

main

main-doc

minicpm-challenge

news_update

perf/cosyvoice3-hopper-default

perf/cosyvoice3-streaming-hift

perf/minimax-h3-adaln-cache-offload

perf/minimax-h3-lossless-hot-path

perf/moss-local-batched-execution

perf/qwen3-tts-event-driven-default

pipeline-npu

pq_dev

pr-6169

qby_dev/cpu_support

qwen-image-2.1

raw-image-logo

readme-update-supported-models

release/v0.12.0rc1

release/v0.14.0

release/v0.14.0.post1

release/v0.14.0rc1

release/v0.15.0rc1

release/v0.16.0

release/v0.16.0rc1

release/v0.17.0rc1

release/v0.18.0

release/v0.18.0.post1

release/v0.18.0rc1

release/v0.19.0rc1

release/v0.20.0

release/v0.21.0rc1

release/v0.22.0

release/v0.22.0.post1

release/v0.22.0rc1

release/v0.23.0rc1

release/v0.24.0

release/v0.24.0rc1

release/v0.24.1

release/v0.25.0rc1

release/v0.26.0

release/v0.26.0rc1

release/v0.27.0rc1

release/v0.28.0

release/v0.28.0rc1

release/v0.29.0rc1

release/v0.30.0

release/v0.30.0rc1

release/v0.31.0rc1

revert-1255-er-stable-audio-online

revert-1338-refine-diffusion-loader

revert-1431-feature/pr-reviewer-auto-trigger

revert-2565-fix/ssrf-image-video-url

revert-3118-voxcpm2-instructions-cfg

revert-3304-FusedRMSNorm

revert-3943-feat/sensenova-u1-fp8-online-mot-gen

revert-4958-perf/code-predictor-fuse-qkv-gateup-v2

revert-5464-test_minicpmo_4_5

revert-5478-BugFixForOmniGPUWorkerBase

revert-6307-codex/6297-topology-aware-diffusion-worker-title

revert-6574-revert-6557-fix/wan-dmd-alignment-test

revert-6624-fix/hf-snapshot-name-match

revert-954-handler_fix

rl-http-interfaces

split-ode-solver

test-ready-label

test/dots-tts-e2e-coverage

test/logprobs-sched-026-mocks

update-committers

update-doc

update-doc-to-014

update-v0200

v0.18.0-release-news

v016

worktree-enable-twine-upload

wzliu_connector_dev

ywang96-patch-1

#4

#6

v0.11.0rc1

v0.12.0rc1

v0.14.0

v0.14.0rc1

v0.15.0rc1

v0.16.0

v0.16.0rc1

v0.17.0rc1

v0.18.0

v0.18.0rc1

v0.19.0rc1

v0.20.0

v0.20.0rc1

v0.21.0rc1

v0.21.0rc2

v0.22.0

v0.22.0rc1

v0.23.0rc1

v0.24.0

v0.24.0rc1

v0.24.1

v0.25.0rc1

v0.26.0

v0.26.0rc1

v0.27.0rc1

v0.28.0

v0.28.0rc1

v0.29.0rc1

v0.30.0

v0.30.0rc1

v0.31.0rc1

Mono Color

• e37c5aed40 probe(evidence): round two - W8A8 int8-dynamic passes the structure gate at BF16-parity PSNR; verdict flips to YES; upstream Marlin report draft staged ISSUE-5 admin 2026-10-08 09:06:49 -04:00
• 704b8e29ae probe(round two): W8A8 int8-dynamic scheme — kernel isolation, raw-op probe, artifact builder support admin 2026-10-08 08:58:45 -04:00
• bf65a41ceb probe(evidence): int8-wo capacity matrix, structure gate, PSNR, Marlin 8-bit NaN isolation (issue 5) admin 2026-10-08 08:00:51 -04:00
• 74b596cd05 probe: artifact round-trip verifier (unpack+dequant vs sealed BF16) admin 2026-10-08 07:45:20 -04:00
• 67e177860a probe: track issue-5 fixtures.json (ignore exception, matches issue-2 kit) admin 2026-10-08 07:39:55 -04:00
• 321e8bb47f fix(quant): load offline-quantized compressed-tensors checkpoints directly on CPU under offload admin 2026-10-08 07:36:01 -04:00
• e778d21f91 probe: int8 weight-only compressed-tensors DiT artifact kit (issue 5) admin 2026-10-08 07:00:16 -04:00
• 2f8363c5fc fix(quant): resolve hyphenated registry methods and disk-dict from_config in quant factory admin 2026-10-08 07:00:05 -04:00
• 4f5f99d8c6 probe(review): fp8 serve row + 3-ref fixture + fp8 quality-collapse diagnosis ISSUE-2 admin 2026-10-08 06:27:36 -04:00
• 67eb7fdb48 probe: Qwen-Image conditioning-KV spill probe on RTX 3090 / RTX 5080 (issue 2) admin 2026-10-08 00:57:51 -04:00
• 3e705fcdb1 [Model] Add JoyAI-Image-Edit pipeline support (#4112) main hurukawa 2026-10-08 08:40:55 +08:00
• 6a70bda0f6 [Bugfix][Qwen3-Omni] Default the talker voice to Ethan (#8568) Yueqian Lin 2026-10-07 17:07:53 -07:00
• 1c71a2c03a [Bugfix] Report sample rate and channels for streamed speech PCM (#8567) Yueqian Lin 2026-10-07 17:07:49 -07:00
• c32a26a845 Add XPU deploy overlays and mark Intel GPU support (#8409) 3 people 2026-10-07 15:19:07 -07:00
• 8484d85d1f [Bugfix] Fix Sender Address Propagation (#8599) Alex Brooks 2026-10-07 15:04:01 -06:00
• 744c202d86 [BugFix] Optimize turn-based /v1/realtime auto truncation (#8566) Nick CaoandCodex 2026-10-07 15:57:09 -04:00
• f3743ebdb6 [Bugfix] Fix In Place Compile Overwrites for Qwen3Omni Embeddings with TP > 1 (#8532) Alex Brooks 2026-10-07 11:01:45 -06:00
• 3345fb7498 [CI/Build] Fix CPU groups in distributed VAE test mocks (#8591) amy-why-3459 2026-10-08 00:24:15 +08:00
• 32ead8e5c8 [NPU] upgrade to v0.31.0 (#8588) Weiming Liao 2026-10-07 23:35:09 +08:00
• c8e69c08fb [CI/Build][ROCm] Stabilize R2-01 platform selection (#8561) haic0andandyluo7 2026-10-07 08:03:03 -07:00
• e54e185593 [feature] Reduce memory usage with SP (#7728) MaciejBalaNV 2026-10-07 16:36:32 +02:00
• 3dc35694b3 [New model] Qwen-Image-2.1 support (rebase + review follow-ups) (#8099) 6 people 2026-10-07 21:23:59 +08:00
• 8ef77be66e [Model] Avoid repeated attention and MLP work in Klein TP (#8553) QianCyrusandHongsheng Liu 2026-10-07 22:21:05 +09:00
• b21df3bcb6 [Frontend] Share regular and duplex chat initialization (#7646) NATURE 2026-10-07 18:49:34 +08:00
• de253412b0 [Core] Defer generation payload host waits to output consumption (#8540) amy-why-3459 2026-10-07 17:20:54 +08:00
• d5a3380103 [CI/Build][ROCm] Run diffusion CPU-offload coverage with SDPA (#8570) haic0andCursor 2026-10-06 19:45:52 -07:00
• 9482bf1e8c [Model] MiniCPM-o 4.5: enable fused CFM body by default on CUDA (#8557) amy-why-3459 2026-10-07 10:25:39 +08:00
• 712b90de9c [Perf][Qwen3-Omni] Capture the thinker image encoder with CUDA graphs (part of #7452) (#8215) 4 people 2026-10-07 09:29:59 +08:00
• ed6bc5ae0c [Perf][PersonaPlex] Batch live duplex appends and replay Mimi codec frames from CUDA graphs (#8356) Yueqian Lin 2026-10-06 17:05:08 -07:00
• e61cd4290f [CI/Build][ROCm] Move Pi0.5 CPU coverage to AMD nightly (#8559) 3 people 2026-10-06 11:05:14 -07:00
• a0d915254b [CI][ROCm] Add and stabilize HunyuanImage3 nightly coverage (#7934) 4 people 2026-10-06 09:54:27 -07:00
• bad88bca25 [Frontend] MiniMax-H3: Add latent-mask editing to the ComfyUI extension (#7575) 3 people 2026-10-07 00:32:40 +08:00
• 7be49aa0ce [Model] Default MOSS Local 1.5 to adaptive CUDA MRV2 with MPS (#8438) Sy03 2026-10-06 08:20:46 -07:00
• ee9ab9ea88 [Perf] Fuse AdaLayerNorm CUDA eager path with Triton (#8231) Dong1017andamy-why-3459 2026-10-06 23:11:45 +08:00
• 182b77da4c [Perf][MiniCPM-o] Stage 0 streaming audio/vision encoder CUDA graphs and candidate sampling (#8430) BeatSeatandamy-why-3459 2026-10-06 22:53:03 +08:00
• 65699fe94c [Core] Generalized turn-based /v1/realtime implementation (#8339) Nick CaoandCodex 2026-10-06 10:07:31 -04:00
• 25fb257667 [Model] MiniMax-H3: accept raw video masks and resolve the latent grid server-side (#7947) 3 people 2026-10-06 20:46:35 +08:00
• 0677064b87 [Bugfix] Migrate CreateAudio from deprecated class Config to ConfigDict (#5245) Nick CaoandCodex 2026-10-06 08:42:09 -04:00
• 5f95115e70 Kandinsky6 (#8537) 3 people 2026-10-06 10:55:15 +03:00
• 1f63e97a40 [CI][ROCm] Add Cosmos3 nightly coverage (#7933) haic0andCursor 2026-10-06 00:18:44 -07:00
• c2cd13c009 [Core] Add batch parallel VAE decode (#8162) QianCyrus 2026-10-06 14:51:35 +09:00
• 2bcd1e449a [CI][ROCm] Add LingBot nightly coverage (#7935) haic0andCursor 2026-10-05 22:31:51 -07:00
• 5e34e1c97d [XPU][Docker] Align XPU image and CI with vLLM v0.31.0 (#8528) Joshna-Medisetty 2026-10-05 22:07:43 -07:00
• 0ad67f1d8f [CI/Build][ROCm] Retry AMD Simple Diffusion shards on GPU hangs (#8520) 3 people 2026-10-05 21:01:31 -07:00
• 099f9b553d [Model] Optimize Wan VAE first-frame decoding with Conv2d (#8179) QianCyrus 2026-10-06 12:44:34 +09:00
• e4a8b2e858 [CI/Build][ROCm] Reuse AITER cache for Mammoth attention tests (#8527) haic0andCursor 2026-10-05 18:35:01 -07:00
• cc444b8875 [Bugfix][MiniCPM-o] Take the duplex vision tile from the image-processor config (#8462) Tianyao Wuandamy-why-3459 2026-10-06 09:04:21 +08:00
• aad3090877 [Model][PersonaPlex] Vectorize prefill embedding construction (#7481) Linze Shi 2026-10-05 11:52:41 -07:00
• ae430dddc0 [Model][PersonaPlex] Encode only consumed Mimi codebooks (#7399) LOGO127 2026-10-06 02:49:33 +08:00
• 79562e97ad [Perf][MiniCPM-o] Code2Wav resident Euler slot pool, fused DiT body, and exact HiFT graphs (#8443) BeatSeatandamy-why-3459 2026-10-06 02:36:28 +08:00
• 3aab12851f [Bugfix] Configure logging in the Omni CLI/API Server (#8518) Nick CaoandCodex 2026-10-05 13:53:14 -04:00
• 539c533762 [Bugfix] Keep client error status in the single-GPU diffusion executor (#8474) Wenbo Ji　嵇文博 2026-10-05 19:09:34 +02:00
• 51b0184969 [Tests] Add tiny model test for StableDiffusionXLPipeline pr-6169 Nick CaoandClaude 2026-08-13 09:24:41 -04:00
• 9146284c16 [Perf] Add CUDA graphs for MiniCPM-o 4.5 input encoders (#8332) amy-why-3459 2026-10-05 23:01:07 +08:00
• a0e626caaf [Bugfix] Validate sleep stage_ids before blocking admission (#8501) Yufeng He 2026-10-05 22:27:53 +08:00
• f3d0dbe5a8 [Bugfix] Preserve host transport fallback when discovery cleanup fails codex/shared-stage-payload-transport Sy03 2026-10-05 22:07:14 +08:00
• 06f562cb3d [Core] Cache host ring discovery without blocking empty polls Sy03 2026-10-05 21:56:29 +08:00
• d98feb3c65 [Bugfix] Claim fallback stage segments before tracking readers Sy03 2026-10-05 12:26:01 +08:00
• 3322a78856 [Core] Optimize mixed stage payloads and scalar restoration Sy03 2026-10-05 11:54:38 +08:00
• bb3f5b5cb7 [Core] Preserve native scalar encoding in stage tensor frames Sy03 2026-10-05 05:12:14 +08:00
• 4f99090212 [Doc] Specify the host ring process namespace requirement Sy03 2026-10-05 04:26:59 +08:00
• 06202d8f2a [Core] Reduce local stage payload overhead with shared host rings Sy03 2026-10-05 04:06:23 +08:00
• 023716ecc4 [Bugfix] Do not assume a stream decoder in MRv2 eager-MTP preprocess (#8484) Sy03 2026-10-05 06:29:38 -07:00
• 5351cd6645 [Bugfix] Qwen3-Omni thinker: create deepstack buffers on the real device (#8461) 3 people 2026-10-05 15:00:04 +02:00
• 091b256674 [Model][Perf] IndexTTS2: drop per-step hidden D2H in latent mode (#8478) ztyan 2026-10-05 20:20:25 +08:00
• 6dd0d1f931 [Bugfix] Update routed-expert tests for native auxiliary output (#8500) v0.31.0rc1 release/v0.31.0rc1 Yueqian Lin 2026-10-05 02:40:07 -07:00
• 67d21b5fff [Core] Add token_ids output type and enforce canonical names (#7843) JungleandCodex 2026-10-05 16:49:03 +08:00
• 0b0d2d7009 [Rebase] Align vLLM-Omni with vLLM v0.31.0 (#8459) Zhou Taichang 2026-10-05 14:11:01 +08:00
• 1f47fa2f29 [Model][Core] Avoid unused CosyVoice3 payloads and redundant AR output work (#8473) Scott Ye 2026-10-05 12:52:19 +08:00
• 07326d0922 Align source installation documentation with vLLM 0.31.0 dev/vllm-align tzhouam 2026-10-05 11:52:40 +08:00
• df14622fb4 Drop CosyVoice regression changes superseded by main tzhouam 2026-10-05 11:41:24 +08:00
• 5ca23106fd [Model] Add SheetSage2 preprocessing and YuE2 score handoff (#8371) 汪志鹏 2026-10-05 11:09:37 +08:00
• bbd539d2b6 [feature] Enable multi-node inference through Ray backend (#7938) 3 people 2026-10-05 04:53:36 +02:00
• c1e84ce946 [Bugfix] Wire output buffers into duplex model test harnesses (#8489) Yueqian Lin 2026-10-04 16:55:19 -07:00
• f781f7bb08 [Core][Frontend] Bound duplex output delivery and reject cancelled audio (#7644) bcsdhjew 2026-10-05 05:06:04 +08:00
• 3f20252151 [Model] Fuse MammothModa2 QK norm and RoPE (#7969) SunandOpenAI Codex 2026-10-05 02:44:45 +08:00
• 855254d4af [Bugfix] Validate MammothModa2 text-to-image dimensions (#7482) TonybotNi 2026-10-05 02:43:35 +08:00
• 65f5a1d5b3 [Bugfix][Examples][MammothModa2] Serve T2I via LLM-typed DiT stage (#7199) + benchmark harness and profiler wiring (#7293) quan-yi-aiandHongsheng Liu 2026-10-05 02:42:58 +08:00
• b9cf74352a [Perf] Support continuous batching for MammothModa2 (#7954) hurukawa 2026-10-05 02:22:37 +08:00
• c9c478dd07 [Perf] Support TeaCache for MammothModa2 (#5357) hurukawa 2026-10-05 01:25:56 +08:00
• c9bee166fb ci: route AR paged attention tests to GPU lanes (#8346) andyluo7 2026-10-04 03:13:21 -07:00
• d2873b38b7 Keep baseline ASR evaluation unchanged and retain documentation WAVs debug/qwen3-omni-doc-audio-artifacts tzhouam 2026-10-04 17:56:18 +08:00
• 5b9032e335 Keep baseline TTS ASR evaluation unchanged and retain failed-audio evidence debug/qwen3-tts-failed-speech-audio tzhouam 2026-10-04 17:56:18 +08:00
• 2f17684774 Preserve baseline ASR fixtures for matching-main failure classification tzhouam 2026-10-04 17:54:28 +08:00
• 591f013d63 Recheck narrowband TTS content with stronger ASR tzhouam 2026-10-04 17:46:32 +08:00
• 0e5e37bde6 Recheck narrowband TTS content with stronger ASR tzhouam 2026-10-04 17:46:32 +08:00
• 4934741df9 [Bugfix] Add TRTLLM attention custom op and execution contract (#7214) Rahul SteigerandRahul Steiger 2026-10-04 11:33:51 +02:00
• e83dd95221 Import regression-test context manager on Omni main tzhouam 2026-10-04 17:15:33 +08:00
• ad7f5f9480 Use accurate ASR for Qwen3-Omni documentation audio tzhouam 2026-10-04 17:12:01 +08:00
• 223e521320 Use accurate ASR for Qwen3-Omni documentation audio tzhouam 2026-10-04 17:12:01 +08:00
• 39b69624c4 Format failed-speech diagnostic helper and regression coverage tzhouam 2026-10-04 16:44:42 +08:00
• e81ae8f66f Format failed-speech diagnostic helper and regression coverage tzhouam 2026-10-04 16:44:42 +08:00
• b9f77fdfbd Retain failed speech ASR input without changing accuracy assertions tzhouam 2026-10-04 16:40:36 +08:00
• 69b2fcb94e Retain failed speech ASR input without changing accuracy assertions tzhouam 2026-10-04 16:40:00 +08:00
• 1e27c8f967 Retain Qwen3-Omni documentation WAVs for failure diagnosis tzhouam 2026-10-04 16:08:23 +08:00
• c942b3b666 Merge current main and retain CosyVoice accumulation regression coverage debug/cosyvoice-conditioning tzhouam 2026-10-04 15:44:54 +08:00
• b65f97bc48 [Feature] Add MiniMax H3 res_multistep sampler (#8378) nodeeeeee 2026-10-04 15:26:19 +08:00
• 148cb943ce [Bugfix][CI/Build] Fix layerwise offload shutdown and enable it for the LTX-2 distilled T2V accuracy test (#8453) YAN YIXINandClaude Opus 5.5 2026-10-04 15:25:51 +08:00
• f7d8d36353 Retain Qwen3-Omni documentation audio for failure diagnosis tzhouam 2026-10-04 15:25:06 +08:00
• 3ab726c126 test: bound CPU vocoder reference threads (#8347) andyluo7 2026-10-03 23:19:22 -07:00