how to train inspire policy

[zyk762]

hello,i want to ask the train_diffusion_policy
i already have dataset like this

and i am training the policy,but it looks like the shape doesn't match,and i notice the difference between your sample data,which is from X_hand and my inspire hand shows like this and it comes with the problem

how can i fix this

looking forward to your reply

[zyk762]

i change the action_dim and it look like work well

[zyk762]

i change the action_dim and it look like work well
it is in train_diffusion_policy.yaml