Returns a demo (list of observations)

Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_mask (

r""" Checks if the input is a valid image. A valid image can be: - A `PIL.Image.Image`. - A 2D or 3D `np.ndarray` or `torch.Tensor` (

r""" Checks if the input is a valid image or list of images. The input can be one of the following formats: - A 4D tensor or numpy array

Launch workers with a consistent spawn strategy. Args: args: Parsed argparse.Namespace with at least worker_num and api_key. targ

r""" Load pretrained attention processor layers into [`UNet2DConditionModel`]. Attention processor layers have to be defined in

Loads or creates a model card. Args: repo_id_or_path (`str`): The repo id (e.g., "runwayml/stable-diffusion-v1-5") or lo

r""" Normalize an image array to [-1,1]. Args: images (`np.ndarray` or `torch.Tensor`): The image array t

Convert a numpy image or a batch of images to a PIL image.

Populates the `model_card` with library name and optional tags.

Concatenate x and its positional encodings, following NeRF. Reference: https://arxiv.org/pdf/2210.04628.pdf

r""" Convert a PyTorch tensor to a NumPy image. Args: images (`torch.Tensor`): The PyTorch tensor to conv

r""" Removes all hooks that were added when using `enable_sequential_cpu_offload` or `enable_model_cpu_offload`.

Updates paths inside attentions to the new naming scheme (local renaming)

r""" Resizes and crops a tensor of images to the specified dimensions. Args: samples (`torch.Tensor`): A

resize the generated frames to the desired size according to the type of the input frames

Best-effort chmod that ignores PermissionError when you are not the owner.

Best-effort chmod that ignores PermissionError when you are not the owner.

Save action sequences as JSON files in the specified directories. Input: - b_action: Int32[Tensor, "b 14"] actions corresponding to

Save a model and its configuration file to a specified directory, allowing it to be re-loaded with the `[`~models.adapter.MultiAdapte

return: output_dict = { "save_dirs": save_dirs, # same with input "pred_frames": pred_frames

Sets a specific adapter by forcing the model to only use that adapter and disables the other adapters. If you are not familiar with

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

Set resolution. Args: x_res (`int`): x resolution of spectrogram (time). y_res (`int`):

r""" Set the switch for the npu flash attention.

Updates the point cloud of the object, and recomputes the position.

Clips values to the range [min_value, max_value]. First checks if any values are out of range and prints a message. Then clips all values to the