Version 3.4.2

Fix:

• Downgrade scikit-image version to avoid reloading the kernel in Colab notebooks
• I2I workflow inference edge case fixed when multiple images in multiple directories where found leading to an infinite loop.

Full Changelog: v3.4.1...v3.4.2

Version 3.4.1

Minor:

• Add BiaPy version into YAML files

Fix:

• Update pyproject.toml specifiying a few libraries more to not be that restrictive and so BiaPy is smoothly installed in Colab

Full Changelog: v3.4.0...v3.4.1

Version 3.4.0

Major:

• Allow to train with large Zarr/H5 files without loading data in memory
• Adapt by_chunks to multichannel data
• Add option to reuse predictions in test

Minor:

• Calculate detection metrics in by_chunks setting
• Add DATA.NORMALIZATION.CUSTOM_MODE option
• Add percentile normalization

Fix:

• Correct bug when discading samples from training when using TRAIN.MINIMUM_FOREGROUND_PER
• Fix PSNR calculation in I2I and SR workflows
• Update notebooks and correct errors related to gdown package
• Correct validation paths in generators when in memory is False

Full Changelog: v3.3.14...v3.4.0

Version 3.3.15

Fix:

• Preprocessing checking
• PSNR calculation in I2I and SR workflows

Full Changelog: v3.3.14...v3.3.15

Version 3.3.14

New functionality:

• Add multithread detection

Fix:

• dense contour creation in 3D
• Add grid masking to MAE
• Ensure validation logs prints metric/loss in the same order as with training data
• Solve problems in recently added Image to Image (I2I) workflow
• Correct TTA errors produced by recent changes when added multihead models

Full Changelog: v3.3.13...v3.3.14

Version 3.3.13

Fix:

• Fix normalization issue during test image reconstruction in SSL and SR workflows

Full Changelog: v3.3.11...v3.3.12

Version 3.3.12

Fix:

• Fix semantic segmentation multiclass edge error when a batch is compose by just one sample
• Correct minor error in SSL workflow using masking (MAE) during inference
• Fix some SR scale problems due to recent changes

Full Changelog: v3.3.11...v3.3.12

Version 3.3.11

Fix:

• Correct few scale problems due to recent changes in SR workflow

Full Changelog: v3.3.10...v3.3.11

Version 3.3.10

Fixes:

• Fix typo in base_workflow.py due to recent changes that crash during inference
• Convert PROBLEM.SUPER_RESOLUTION.UPSCALING to a tuple to use one per dimension
• Add TRAIN.OPT_BETAS
• Change TRAIN.W_DECAY default value to 0.02
• Fix bug when in DATA.TRAIN.REPLICATE
• Correct 3D SR workflow

Full Changelog: v3.3.9...v3.3.10

Version 3.3.9

Quick patch:

• Fix identation error in some models.

Full Changelog: v3.3.8...v3.3.9