delu#

Deep Learning Utilities for PyTorch users.

Tensor operations

`to`	Like `torch.Tensor.to`, but for collections of tensors.
`cat`	Like `torch.cat`, but for collections of tensors.
`iter_batches`	Iterate over tensor or collection of tensors by (random) batches.

Tools

`EarlyStopping`	Performs early stopping after N consequtive bad (non-improving) updates.
`Timer`	Measures time.

Submodules

See the sidebar on the left.

Warning

The API listed below is deprecated.

`ProgressTracker`	[DEPRECATED \| Instead, use `delu.EarlyStopping`.
`Stream`	[DEPRECATED \| Currently, no alternatives are available.] Smart wrapper for data loaders and iterables.

`collate`	[DEPRECATED \| Instead, use `torch.utils.data.default_collate`] Almost an alias for `torch.utils.data.dataloader.default_collate`.
`concat`	[DEPRECATED \| Instead, use `delu.cat`]
`improve_reproducibility`	[DEPRECATED \| Instead, use `delu.random.seed` and manually set flags mentioned in the PyTorch docs on reproducibility] Set seeds and turn off non-deterministic algorithms.

[DEPRECATED | Instead, use model.eval() + torch.no_inference/no_grad] Context-manager & decorator for models evaluation.