modelzoo.common.pytorch.utils#

General purpose Pytorch Utilities

Functions

get_checkpoints

Gather checkpoints in a model directory

is_mup_run

Check if the run is configured with muP hyperparameter settings

setup_artifact_dir

Create a unique subdirectory for this run by generating a time stamp so that parallel runs using the same model_dir don't overwrite common files.

setup_logging

Configure default logging format

to_cpu

Move tensor from device to cpu

to_tensor

If the provided value is a Python int or float, it converts them into PyTorch Tensors of type int32 and float32 respectively.

visit_structure

Recursively traverse nested structure and return the items accepted by the selector.

Classes

BufferedShuffleDataset

Dataset shuffled from the original dataset.

IterableDatasetSampler

This sampler can be used with a multi-worker distributed dataloader.

RunConfigParamsValidator

Validate Run Configs

SampleGenerator

Iterator which returns multiple samples of a given input data.