Write text to stdout (and a file) and optionally flush.

obj_filename: path to obj file gif_dst_path: if set a path, will render n_views frames, then save it to a gif file

seed everthing

Traverses the object name and returns the last (rightmost) python object.

@param: H is homography matrix of dimension (3x3) @param: pt is the (x, y) point to be transformed Return: returns a transfo

Additional filtering is applied to filter out-of-box samples. Modifications made by Zexin He.

args: prompr: str seed: int steps: int return: rgb: PIL.Image

can set views_pil, cond_pil simutaously or set gif_pil only seed: int target_face_count: int save_fo

Evaluate a function without caching intermediate activations, allowing for reduced memory at the expense of extra compute in the backward pas

Flush written text to both stdout and a file, if open.

this will perform the index select, cat the tensors, and provide the attn_bias from cache

Searches for the underlying module behind the name to some python object. Returns the module and the object name (original name with module part r

Return depths of approximately uniformly spaced samples along rays.

Performs a single optimization step. Args: closure (callable, optional): A closure that reevaluates the model and

convert a PIL.Image to rgb mode with white background maybe_rgba: PIL.Image return: PIL.Image

Fast CUDA implementation of `bias_act()` using custom ops.

Slow reference implementation of `bias_act()` using standard TensorFlow ops.

Finds the python object with the given name and calls it as a function.

Flush, close possible files, and remove stdout/stderr mirroring.

DINOv2 ViT-B/14 model (optionally) pretrained on the LVD-142M dataset.