Update vllm_omni/core/sched/diffusion_scheduler.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Zhou Taichang <tzhouam@connect.ust.hk>
This commit is contained in:
Zhou TaichangandCopilot authored and GitHub committed 2025-10-22 17:13:45 +08:00
1 parent 19220aae82
commit 8d6a360f96
1 file changed
+3 -3
+3 -3
View File
@@ -11,9 +11,9 @@ from vllm_omni.core.sched.output import OmniNewRequestData
class DiffusionScheduler(OmniScheduler):
def schedule(self) -> SchedulerOutput:
"""扩散快速通道:
- 一次性喂入该请求的全部输入 token(若为 0,则分配 1 个占位 token)。
- 若无法一次性满足 token 预算,则退回上游 vLLM 的默认调度。
"""Diffusion fast path:
- Feed all input tokens of the request at once (if 0, allocate 1 placeholder token).
- If the token budget cannot be satisfied at once, fall back to the default vLLM scheduling.
"""
# 选出零 prompt 且使用 pooling(扩散结果经 pooler_output 回传)的请求