diag: F 段——MeanFlow 一步采样残留噪声（开环 1/2/4/8 段与零噪声的幅度/抖动/残差-噪声相关；闭环 4 段、零噪声）；sample_k 加 steps、rollout 加 steps/z_scale；删 E 段；本机执行清单更新

ml/sim/closed_loop.py

@@ -65,8 +65,8 @@ def dls_steps(kin, q, T, R, n=3, scale=0.3, lam=0.5):
 
 
 @torch.no_grad()
-def rollout(model, kin, tasks, device, K=8, safety=False, seed=0, max_sec=None, return_traj=False, use_sub=True, sub_override=None, qp0=None):
-    """use_sub=False：不用子目标（sf=0）；sub_override=(subs [B,2,6] 绝对关节角, need [B] bool)：冷启动时用给定途经点代替子目标头；qp0 [B,6]：起点的上一帧（默认 = q0，即静止起步）。均为诊断用"""
+def rollout(model, kin, tasks, device, K=8, safety=False, seed=0, max_sec=None, return_traj=False, use_sub=True, sub_override=None, qp0=None, steps=1, z_scale=1.0):
+    """use_sub=False：不用子目标（sf=0）；sub_override=(subs [B,2,6] 绝对关节角, need [B] bool)：冷启动时用给定途经点代替子目标头；qp0 [B,6]：起点的上一帧（默认 = q0，即静止起步）。steps：MeanFlow 采样段数；z_scale：噪声缩放（0 = 只取「均值」样本）。均为诊断用"""
     g = torch.Generator(device=device); g.manual_seed(seed)
     q0, T, R, ex, tr = tasks_to_tensors(tasks, device)
     B = q0.shape[0]
@@ -102,8 +102,8 @@ def rollout(model, kin, tasks, device, K=8, safety=False, seed=0, max_sec=None,
                 nsub = torch.where(adv, nsub - 1, nsub); sub_t = torch.where(adv, torch.full_like(sub_t, tnow), sub_t)
             sf = (nsub > 0).float()
             feat = features(kin, q, qp, T, R, pre, torch.full((B,), r, device=device), subs[:, 0], sf)
-            z = torch.randn(B * K, H, 6, device=device, generator=g)
-            act, val, sub = model.sample_k(feat, K, z)
+            z = torch.randn(B * K, H, 6, device=device, generator=g) * z_scale
+            act, val, sub = model.sample_k(feat, K, z, steps=steps)
             inc = act.clamp(-1, 1) * step
             if cold:
                 sub0 = sub