Helper function to compute the semantic, instance, panoptic segmentation metrics of a model on a S3DIS 6-fold, for given instance graph and pa

compute per frame sdf

Helper function to compute the semantic segmentation metrics of a model on a S3DIS 6-fold. :param fold_ckpt: dict Dictionary with S3D

Fusion all frames, perform adaptive sdf_funcation on the contract spaces.

Move the CSRData to the CPU.

Move the NAG with all Data in it to CUDA.

Move the CSRData to the first available GPU.

Returns a directory path in root, named based on the current date and time.

Extend `torch_geometric.Data.detach` to handle Cluster and InstanceData attributes.

Detach all tensors in the CSRBatch.

Device of the first-encountered tensor in 'self'.

Download the content of an URL to a specific folder. :param url: string :param folder: string :param log: bool If `False`, will n

Simply applies a sigmoid on `edge_affinity_logits` to produce the actual affinity predictions to be used for superpoint graph clusteri

All keys starting with `edge_`, apart from `edge_index` and `edge_attr`.

Prompts user to input tags from command line if no tags are provided in config.

Extracts a flattened list of transforms from a Compose or from a list of transforms.

Same as torch.zeros but relies numpy on CPU. This may be x40 faster when manipulating large tensors on CPU. ``` from time import time

filter out small groups of points

Heuristic: keeping ```f_min = 1 / f_max``` ensures that roughly 50% of the encoding dimensions are untouched and free to use. This is imp

Panoptic predictions on the full-resolution input point cloud. Returns the predicted semantic label and instance index for each point,

Semantic predictions on the full-resolution input point cloud. Final semantic segmentation predictions are the argmax of the

Garbage collection Torch (CUDA) memory.

Generate multiple images with random ground truth and predicted instance and semantic segmentation data. To make the images realistic, and to

Helps `self.from_list()` and `self.to_list()` identify which classes to use for batch collation and un-collation.

Helps `self.from_list()` and `self.to_list()` identify which classes to use for batch collation and un-collation.

Helps `self.from_list()` and `self.to_list()` identify which classes to use for batch collation and un-collation.

Safely retrieves value of the metric logged in LightningModule.

Initializes multi-GPU-friendly python command line logger.