MickandClaude Fable 5 754b692afc [diffusion] optimization: support cuda-ipc zero-staging all-to-all for 2-rank Ulysses (#31854) ...

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

2026-07-31 19:35:48 +08:00

..

breakable_cuda_graph

Move breakable CUDA graph back into model_executor/runner_backend_utils (#30586)

2026-07-09 14:19:16 +08:00

cache

[diffusion][cache-dit] add dual-transformer Cache-DiT adapter specs (#30150)

2026-07-07 08:10:52 -07:00

disaggregation

[diffusion] disagg: handle numpy arrays in cross-role transfer field extraction (#31325)

2026-07-20 00:50:17 -07:00

distributed

[diffusion] optimization: support cuda-ipc zero-staging all-to-all for 2-rank Ulysses (#31854)

2026-07-31 19:35:48 +08:00

entrypoints

[Diffusion] Return scheduler sigmas snapshot in rollout dit_trajectory (#32683)

2026-07-31 00:29:07 -07:00

layers

[diffusion] optimization: support cuda-ipc zero-staging all-to-all for 2-rank Ulysses (#31854)

2026-07-31 19:35:48 +08:00

loader

[diffusion] feat: unify encoder folding and batch data-parallel encoding (#30211)

2026-07-30 20:15:22 +08:00

managers

[diffusion] optimization: support cuda-ipc zero-staging all-to-all for 2-rank Ulysses (#31854)

2026-07-31 19:35:48 +08:00

models

[diffusion] optimization: support cuda-ipc zero-staging all-to-all for 2-rank Ulysses (#31854)

2026-07-31 19:35:48 +08:00

pipelines

[diffusion] post_training: Add LoRA IPC weight sync via lora_merge mode (#31029)

2026-07-15 08:42:54 +08:00

pipelines_core

[diffusion] feat: unify encoder folding and batch data-parallel encoding (#30211)

2026-07-30 20:15:22 +08:00

platforms

[diffusion] refactor: remove stale kernels and dead code (#32651)

2026-07-29 06:23:23 +08:00

post_training

[Diffusion] Return scheduler sigmas snapshot in rollout dit_trajectory (#32683)

2026-07-31 00:29:07 -07:00

postprocess

[diffusion] improve: improve realtime webui playback pacing (#27148)

2026-06-04 00:33:56 +08:00

realtime

[diffusion] refactor: refactor realtime control state and adapters (#27698)

2026-06-10 12:27:16 +08:00

server_args

[diffusion] feat: unify encoder folding and batch data-parallel encoding (#30211)

2026-07-30 20:15:22 +08:00

utils

perf(diffusion): decode Wan VAE in BF16 (#32697)

2026-07-29 21:57:50 +08:00

vla

model: support Pi0.5 (#30633)

2026-07-11 07:50:58 +08:00

ipc_array.py

[diffusion] optimize: optimize frame returns path (#24616)

2026-05-08 12:10:09 +08:00

launch_server.py

model: support Pi0.5 (#30633)

2026-07-11 07:50:58 +08:00

scheduler_client.py

[Diffusion] Diffusion model support log-requests (#23049)

2026-07-05 12:01:26 +03:00

server_warmup.py

[diffusion] refactor: remove stale kernels and dead code (#32651)

2026-07-29 06:23:23 +08:00

warmup_request_builder.py

model: support Pi0.5 (#30633)

2026-07-11 07:50:58 +08:00