Return Euler angles from rotation matrix for specified axis sequence. Note that many Euler angle triplets can describe one matrix. Parameter

Normalize, resize, and pad images and stack them into a tensor. Args: observation (dict[str, Tensor]) Returns:

Normalize, resize, and pad images and stack them into a tensor. Args: observation (dict[str, Tensor]) Returns:

If `prompt` is provided, modify it to PaliGemma format and tokenize it. If `lang_tokens` and `lang_masks` are provided, use them directly.

If `prompt` is provided, modify it to PaliGemma format and tokenize it. If `lang_tokens` and `lang_masks` are provided, use them directly.

Convert quaternion to rotation of angle around axis Parameters ---------- quat : 4 element sequence w, x, y, z forming quaternion.

Calculate rotation matrix corresponding to quaternion Parameters ---------- q : 4 element array-like Returns ------- M : (3,

Main entry point.

task_range: [start, end) for splitting tasks

Save model checkpoint in the same format as pretrained model.

Set the scene - Set up the basic scene: light source, viewer.

TODO

r""" TODO

Load training state (optimizer + step) after FSDP and optimizer creation.

Return angle, axis corresponding to Euler angles, axis specification Parameters ---------- ai : float First rotation angle (accor

Return rotation matrix from Euler angles and axis sequence. Parameters ---------- ai : float First rotation angle (according to `

r""" Forward pass through the diffusion model Args: x (List[Tensor]): List of input video tensors, each w

Generates an action with the VLA policy.

TODO: show be rewritten as a dict