Change Log#
Radionets v0.4.0 (2025-08-06)#
API Changes#
Bug Fixes#
replaced variance with standard deviation [#155]
compute uncertainty histogram on source pixels only
check for existing sampling file before generating
proper error propagation after using a normalization [#164]
fix evaluate_msssim_sampled
fix call of evaluate_msssim_sampled [#171]
Fix legend handles missing in eval contour plot [#175]
Data Model Changes#
New Features#
added eval methods for sampled images [#155]
add zenodo.json [#161]
add new normalizing method which normalizes every image
add saving options for more evaluation methods [#165]
Added new block types for deep ResNets and UNets:
BottleneckResBlock,Encoder, andDecoderIntroduced new submodules
activation,archs, andblocksRelated classes (e.g., blocks, activation functions) and functions are moved from other submodules into these respective submodules
Improves readability and reusability
Added a logger that replaces bare prints throughout the code base
Added new diverging colormap radionets.PuOr that is used for plotting in radionets [#178]
Added docs with API references and user/dev guides [#187]
Maintenance#
use mamba in tests [#159]
add MANIFEST.in [#160]
clean up docs/changes [#163]
delete unused functions
delete unsued architectures
rename symmetry function [#166]
add radionets logo to README [#169]
fix comet_ml callback
update process_prediction with better if statement
change hardcoded values for sampling [#170]
set number of bins for histogram plotting [#171]
Update pyproject and python version support
Switch to hatchling build backend [#176]
Restructure dl_framework.architecture [#177]
Flattened module hierarchy
Refactord architecture into modular components, see new features
Moved
LocallyConnected2dclass tounc_archsRefactored some callback submodule
Add error handling for cases where normalization attributes (
self.learn.normalize.mode) may not be definedRemove unnecessary calls to
create_OrBu(), replace it with direct import ofOrBu
Refactored plotting tools and grouped plotting functions into logical groups
CI: Replaced pip with uv for package installation and added codecov test analytics [#178]
Refactoring and Optimization#
add keyword for half of the image
distinguish between tensor and array in get_ifft
fix micromamba installation [#168]
Radionets 0.3.0 (2023-08-04)#
API Changes#
Bug Fixes#
New Features#
Enabled training and evaluation of half sized images (for 128 pixel images) [#140]
Added naming to save path, so that the files do not get overwritten as easily [#144]
Added normalization callback with two different techniques
Updated plotting routines for real/imag images
Updated
evaluate_areaandevaluate_ms_ssimfor half imagesAdded
evaluate_ms_ssimfor sampled images [#146]
Add evaluation of intensity via peak flux and integrated flux comparison [#150]
Centered bin on 1 for histogram evaluation plots
Added color to legend [#151]
Added prettier labels and descriptions to plots [#152]
Maintenance#
Deleted unusable functions for new source types
Deleted unused hardcoded scaling [#140]
Added masked loss functions
Sorted bundles in simulations
Minor adjustments in plotting scripts [#141]
Consistent use of batch_size [#142]
Added the model name to predictions and sampling file
Deleted unnecessary pad_unsqueeze function
Added amp_phase keyword to sample_images
Fixed deprecation warning in sampling.py
Added image size to test_evaluation.py routines [#146]
Outsourced preprocessing steps in
train_inspection.py[#148]Removed unused
norm_pathfrom all instances [#153]Deleted cropping
Updated colorbar label
Removed
source_listargument [#154]
Refactoring and Optimization#
Optimized
evaluation.utils.trunc_rvswith numba, providing functions compiled for cpu and parallel cpu computation. [#143]
Radionets 0.2.0 (2023-01-31)#
API Changes#
Train on half-sized iamges and applying symmetry afterward is a backward incompatible change
Models trained with early versions of
radionetsare not supported anymore [#140]
Bug Fixes#
Fixed sampling of test data set
Fixed same indices for plots [#140]
New Features#
Enabled training and evaluation of half sized images (for 128 pixel images) [#140]
Maintenance#
Deleted unusable functions for new source types
Deleted unused hardcoded scaling [#140]
Refactoring and Optimization#
Radionets 0.1.18 (2023-01-30)#
API Changes#
Bug Fixes#
New Features#
Added creation of uncertainty plots
Changed creation and saving/reading of predictions to
dictsPrediction
dictshave 3 or 4 entries depending on uncertainty
Added scaled option to
get_ifftCreated new dataset class for sampled images
Created option for sampling and saving the whole test dataset
Updated and wrote new tests [#129]