✨ Version 0.5.0

[o-laurent]

New Features

• Introduce TUCheckpoint callbacks, simplifying saving multiple checkpoints for Classification, Segmentation, and Regression routines.

QoL improvements

• lightning<2.5.0 is supported again!
• Add eval_batch_size parameter to all datamodules
• Enable overloading default train and test transforms in datamodules (resolve 🔨Transforms in Datamodules should be settable in the __init__() #163)
• deep_ensembles parameters can be stored on CPU and loaded one model at a time on the GPU during the forward pass

Bug fixes

• Will fix 🐛mc_dropout incompatible with mlp #152.

[codecov]

Codecov Report

Attention: Patch coverage is 91.75109% with 57 lines in your changes missing coverage. Please review.

Project coverage is 98.49%. Comparing base (b0dd7c9) to head (2d9c740).

✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
torch_uncertainty/models/wrappers/zero.py	23.68%	29 Missing ⚠️
torch_uncertainty/datamodules/abstract.py	67.74%	5 Missing and 5 partials ⚠️
..._uncertainty/datamodules/classification/cifar10.py	91.66%	1 Missing and 1 partial ⚠️
...uncertainty/datamodules/classification/cifar100.py	92.00%	1 Missing and 1 partial ⚠️
...uncertainty/datamodules/classification/imagenet.py	92.30%	1 Missing and 1 partial ⚠️
...ch_uncertainty/datamodules/classification/mnist.py	88.88%	1 Missing and 1 partial ⚠️
...tainty/datamodules/classification/tiny_imagenet.py	91.66%	1 Missing and 1 partial ⚠️
torch_uncertainty/layers/packed.py	93.33%	1 Missing and 1 partial ⚠️
...rtainty/post_processing/conformal/conformal_aps.py	96.96%	1 Missing and 1 partial ⚠️
...tainty/post_processing/conformal/conformal_raps.py	97.33%	1 Missing and 1 partial ⚠️
... and 2 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #160      +/-   ##
==========================================
- Coverage   99.15%   98.49%   -0.66%     
==========================================
  Files         144      153       +9     
  Lines        7307     7797     +490     
  Branches      942     1015      +73     
==========================================
+ Hits         7245     7680     +435     
- Misses         28       71      +43     
- Partials       34       46      +12     

Flag	Coverage Δ
pytest	98.49% <91.75%> (-0.66%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.