Release 0.13.0

[BenjaminBossan]

This is the PR to prepare the skorch v0.13 release.

Apart from updating VERSION and CHANGES.md, two deprecations were removed:

• Remove CVSplit: It was renamed to ValidSplit but we kept the old class for backwards compatibility. The deprecation period is over, so remove the class and associated test.
• Remove bc code in get_iterator: In Initialize iterator only once per fit call #835, we made a change to the data loader initialization. To keep backwards compatibility, we had to add extra code to get_iterator, together with a deprecation notice. The deprecation period is over, so that code is removed, as well as the associated test.

Release text

This is the suggested announcement for the release page:

The new skorch release is here and it has some changes that will be exiting for some users.

• First of all, you may have heard of the PyTorch 2.0 release, which includes the option to compile the PyTorch module for better runtime performance. This skorch release allows you to pass compile=True when initializing the net to enable compilation.
• Support for training on multiple GPUs with the help of the accelerate package has been improved by fixing some bugs and providing a dedicated history class. Our documentation contains more information on what to consider when training on multiple GPUs.
• If you have ever been frustrated with your neural net not training properly, you know how hard it can be to discover the underlying issue. Using the new SkorchDoctor class will simplify the diagnosis of underlying issues. Take a look at the accompanying notebook

Apart from that, a few bugs have been fixed and the included notebooks have been updated to properly install requirements on Google Colab.

We are grateful for external contributors, many thanks to:

• Kshiteej K (kshitij12345)
• Muhammad Abdullah (abdulasiraj)
• Royi (RoyiAvital)
• Sawradip Saha (sawradip)
• y10ab1 (y10ab1)

Find below the list of all changes since v0.12.1 below:

Added

• Add support for compiled PyTorch modules using the torch.compile function, introduced in PyTorch 2.0 release, which can greatly improve performance on new GPU architectures; to use it, initialize your net with the compile=True argument, further compilation arguments can be specified using the dunder notation, e.g. compile__dynamic=True
• Add a class DistributedHistory which should be used when training in a multi GPU setting (Add DistributedHistory for multi-gpu training #955)
• SkorchDoctor: A helper class that assists in understanding and debugging the neural net training, see this notebook (skorch doctor: a tool to understand the net #912)
• When using AccelerateMixin, it is now possible to prevent unwrapping of the modules by setting unwrap_after_train=True (Fix issues with unwrapping modules with accelerate #963)

Fixed

• Fixed install command to work with recent changes in Google Colab (Fixed imports and Checked Notebooks #928)
• Fixed a couple of bugs related to using non-default modules and criteria (Bugfixes: Bugs stemming from the use of non-standard modules and criteria #927)
• Fixed a bug when using AccelerateMixin in a multi-GPU setup (Fix accelerate bug, document, add scripts #947)
• _get_param_names returns a list instead of a generator so that subsequent
  error messages return useful information instead of a generator repr
  string (Unhelpful error message when setting invalid parameter with set_params #925)
• Fixed a bug that caused modules to not be sufficiently unwrapped at the end of training when using AccelerateMixin, which could prevent them from being pickleable (Fix issues with unwrapping modules with accelerate #963)

[BenjaminBossan]

Oops, sorry, didn't intend to ping the contributors :)

[RoyiAvital]

@BenjaminBossan , We're honored. We're satisfied users :-).

[sawradip]

Feels good when even small contributions ship to a release, specially as a new contributor.

[thomasjpfan]

Thank you for starting the release process. Likely good to also include #963 and #965. Otherwise, LGTM.

[BenjaminBossan]

Thx @thomasjpfan I included the latest PRs.