[Core] Add UniProc diffusion executor for single-GPU (#6308)

Signed-off-by: Rahul Steiger <rsteiger@nvidia.com>
Signed-off-by: Rahul Steiger <rsteiger@aws-cmh-slurm-1-vscode-04.cm.cluster>
Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.com>
Co-authored-by: Rahul Steiger <rsteiger@aws-cmh-slurm-1-vscode-04.cm.cluster>
This commit is contained in:
authored and GitHub committed 2026-08-22 21:58:21 +08:00
1 parent 72ee535fcd
commit b39f599b09
10 files changed
+619 -60

No files matched your search

+4 -2
View File
@@ -74,8 +74,10 @@ For fields whose deploy default is `null`, the deploy layer contributes no
override. The effective value may still come from a platform section, an
explicit CLI or stage override, or the downstream vLLM engine default.
Note: for diffusion path, `distributed_executor_backend` currently defaults to
`mp`, and `ray` / `external_launcher` are not fully supported yet.
Note: for the diffusion path, an omitted `distributed_executor_backend` selects
`uni` on a single GPU (in-process worker, no MessageQueue / `/dev/shm` output
segments) and `mp` when `num_gpus > 1`. Set `mp` explicitly to keep a worker
subprocess on one GPU. `ray` / `external_launcher` are not fully supported yet.
### Stage fields