admin e778d21f91 probe: int8 weight-only compressed-tensors DiT artifact kit (issue 5) ...

quantize_dit_int8wo.py builds the offline W8A16 artifact (per-channel
symmetric int8, pack-quantized via compressed-tensors' own packer,
per-component config wrapper so encoder/VAE stay BF16, digest manifest);
serve.sh/probe.py/gate.py/psnr.py adapt the issue-2 kit for the
structure-gated capacity matrix (1-4 refs x seeds 2301-2304).

2026-10-08 07:01:55 -04:00

..

assets

[Model] Add YuE2-3B text-to-music (model + /v1/audio/speech adapter) (#7886)

2026-10-01 18:04:31 +08:00

attention

[BugFix][NPU] Fix Fish Speech S2 Pro NPU support (#7546)

2026-09-16 11:38:06 -07:00

benchmarks

[Bugfix] Report sample rate and channels for streamed speech PCM (#8567)

2026-10-07 17:07:49 -07:00

buildkite

[CI/Build][ROCm] Stabilize R2-01 platform selection (#8561)

2026-10-07 08:03:03 -07:00

clients

[Bugfix] Fix Nemotron VoiceChat duplex nightly handshake (#7884)

2026-09-25 17:25:15 +00:00

config

[Model] Default MOSS Local 1.5 to adaptive CUDA MRV2 with MPS (#8438)

2026-10-06 08:20:46 -07:00

core

[Bugfix] Fix Sender Address Propagation (#8599)

2026-10-07 15:04:01 -06:00

dfx

[Rebase] Align vLLM-Omni with vLLM v0.31.0 (#8459)

2026-10-05 14:11:01 +08:00

diffusion

[Model] Add JoyAI-Image-Edit pipeline support (#4112)

2026-10-08 08:40:55 +08:00

distributed

[Rebase] Align vLLM-Omni with vLLM v0.31.0 (#8459)

2026-10-05 14:11:01 +08:00

docs

[Core][Diffusion] Add a typed pre-D2H video media contract (#6615)

2026-09-10 09:50:54 +00:00

e2e

[Model] Add JoyAI-Image-Edit pipeline support (#4112)

2026-10-08 08:40:55 +08:00

engine

[Core] Add token_ids output type and enforce canonical names (#7843)

2026-10-05 16:49:03 +08:00

entrypoints

[Bugfix] Report sample rate and channels for streamed speech PCM (#8567)

2026-10-07 17:07:49 -07:00

examples

[Core] Generalized turn-based /v1/realtime implementation (#8339)

2026-10-06 22:07:31 +08:00

helpers

[Rebase] Align vLLM-Omni with vLLM v0.31.0 (#8459)

2026-10-05 14:11:01 +08:00

host_weight_runtime

[Bugfix][Core] Clean up failed HWR atomic metadata writes (#6956)

2026-09-11 09:03:26 +08:00

inputs

[Rebase] Align vLLM-Omni with vLLM v0.31.0 (#8459)

2026-10-05 14:11:01 +08:00

kernels

feat(magi2): add fused BF16 routed MoE path (#7206)

2026-09-28 23:20:30 +08:00

metrics

[Rebase] Align vLLM-Omni with vLLM v0.31.0 (#8459)

2026-10-05 14:11:01 +08:00

model_executor

[Bugfix][Qwen3-Omni] Default the talker voice to Ethan (#8568)

2026-10-07 17:07:53 -07:00

model_extras

[New model] Qwen-Image-2.1 support (rebase + review follow-ups) (#8099)

2026-10-07 21:23:59 +08:00

model_tests

[Model] Add JoyAI-Image-Edit pipeline support (#4112)

2026-10-08 08:40:55 +08:00

platforms

[CI][ROCm] Add and stabilize HunyuanImage3 nightly coverage (#7934)

2026-10-06 09:54:27 -07:00

profile

[Refactor][CI] Reorganize test tree, and move slow e2e to weekly (#4951)

2026-07-29 23:53:03 +08:00

protocol

[Core][Frontend] Bound duplex output delivery and reject cancelled audio (#7644)

2026-10-04 14:06:04 -07:00

quantization

probe: int8 weight-only compressed-tensors DiT artifact kit (issue 5)

2026-10-08 07:01:55 -04:00

reasoning

[Refactor][CI] Reorganize test tree, and move slow e2e to weekly (#4951)

2026-07-29 23:53:03 +08:00

rl_rollout

[P0][RFC #3747] RL rollout serving: session management + world_model_env (#4770)

2026-09-14 09:26:44 +08:00

tools

[Model] Add SheetSage2 preprocessing and YuE2 score handoff (#8371)

2026-10-05 11:09:37 +08:00

utils

[Frontend] Tune reference cache limits and registered MOSS voice requests (#7883)

2026-09-29 04:10:17 +00:00

worker

[Core] Defer generation payload host waits to output consumption (#8540)

2026-10-07 17:20:54 +08:00

worker_v2

[Model] Default MOSS Local 1.5 to adaptive CUDA MRV2 with MPS (#8438)

2026-10-06 08:20:46 -07:00

__init__.py

[Fix] Resolve comments & update vLLM-Omni name usages. (#122)

2025-12-01 08:05:24 +08:00

conftest.py

test: bound CPU vocoder reference threads (#8347)

2026-10-04 19:19:22 +13:00