atq-v22so3-labelgated-2.0x-60k

atq-v22so3-labelgated-2.0x-60k

ATQ label-gated MoE, 2.0x compressed group, v22 VLM-only conf label, SO(3)-composed merged rotation GT, RoboCasa 60k

60,000 steps, seed 42, RoboCasa 24 tasks, 2 GPUs.

Labels: prehj/robocasa-conf-labels-v22 (VLM only, stride-16 anchors interpolated to every frame; frames outside the anchor span are masked out of the conf loss).

conf_threshold in the config is the training default -- it is an eval-only knob (the head regresses conf and never thresholds it), so override it at serve time with --conf-threshold.

optimizer.pt is not included: the run finished its schedule, so there is nothing to resume; the weights are what you want.

head config

moe_speed = 2.0
moe_label_gated = True
moe_num_experts = 4
conf_threshold = 0.55
action_merge_reduction = sum
action_horizon = 16
discrete_action_dims = [6, 11]