Parameters ---------- name : string Name of the method or property to call. *args Arguments

Parameters ---------- timeout : int or float, optional Number of seconds before the call to `step_wait` times out

Close the environments & clean up the extra resources (processes and pipes). Parameters ---------- timeout : int or f

Close the environments.

for vmap

for vmap

for vmap

for vmap

Discretize the SDE in the form: x_{i+1} = x_i + f_i(x_i) + G_i z_i. Useful for reverse diffusion sampling and probabiliy flow sampling.

Create discretized iteration rules for the reverse diffusion sampler.

DDPM discretization.

x : [ batch_size x in_channels x horizon_steps ] cond : [ batch_size x cond_dim] returns: out : [ batch_size x out_c

x: (B, Ta, act_dim) time: (B,) or int, diffusion step cond: dict with key state/rgb; more recent obs at the end s

x: (B, Ta, Da) time: (B,) or int, diffusion step cond: dict with key state/rgb; more recent obs at the end state:

Modifying denoising schedule

Forward pass for sampling actions. Args: cond: dict with key state/rgb; more recent obs at the end state

Forward pass for sampling actions. Used in evaluating pre-trained/fine-tuned policy. Not modifying diffusion clipping. Args:

assume state-only, no rgb in cond

inputs: x: (B, Ta, Da) time: (B,) floating point in {0,1/2,1/4,1/8,...1/2^n} shortcut flow time cond: dic

inputs: x: (B, Ta, Da) time: (B,) floating point in {0,1/2,1/4,1/8,...1/2^n} shortcut flow time cond: dic