v2.0.0rc2

Release v2.0.0rc2

v2.0.0rc1

[2.0.0rc1] – 2025-07-24

This release marks a major overhaul of ClinicaDL, refactoring the entire framework into a modular, API-first design.

Previously centered around the command line, ClinicaDL now provides a flexible Python API that allows users to build and customize deep learning pipelines with high-level configuration objects. The goal is to make the code more maintainable, scalable, and user-extensible, while preserving the core values of ClinicaDL: reproducibility, robust support for neuroimaging, and data leakage prevention.

This is the first release candidate for version 2.0.0, with the final release planned for September 2025.

Highlights

• Full rewrite of the core library — now fully object-oriented and modular.
• New modules (see next section)
• New MAPS architecture for managing model outputs and metadata.
• Clear configuration-based design with JSON files and dedicated config classes based on pydantic.
• Modern deep learning tooling: PyTorch, MONAI, TorchIO, HuggingFace, MLflow, and Weights & Biases support.
• Extensive and fully updated documentation.

Added

Core Modules:

• Trainer: high-level training controller managing full model lifecycle.
• ClinicaDLModel: flexible base class to define and extend custom architectures.
• CapsDataset: redesigned dataset class in the new dataset module, tailored for CAPS/MAPS.
• Splitter: new module for managing train/val/test split logic.
• Maps: a structured and reproducible representation of model outputs and metadata.

Configuration Classes:

• OptimizationConfig, DataloaderConfig, LossConfig, TransformConfig, MetricConfig, etc.
• Designed to be composable and readable using TOML files.
• Stored alongside results to ensure experiment traceability.

Integration with Modern Tools:

• Transforms: use of torchio and monai.transforms for preprocessing and data augmentation.
• Metrics: integration with MONAI metrics and support for custom metrics.
• Networks: fully compatible with native PyTorch models.
• Logging: support for MLflow and Weights & Biases (W&B) out of the box.
• HuggingFace: integration point for loading pretrained models and tokenizers.

Documentation:

• Fully rewritten Sphinx documentation with improved structure and usage examples.
• Interactive object documentation and visual MAPS structure navigation.

Changed

• All pipelines removed and replaced by a unified API-driven interface.
• Internal architecture redesigned for independent modules that can be combined or extended.
• CLI options replaced by TOML configuration — reducing duplication and increasing clarity.
• All training now done through Trainer, using configuration objects and custom hooks.

Removed

• All legacy CLI commands (e.g., clinicadl train, clinicadl random-search, etc.).
• Old pipelines (train_from_json, preprocessing run, etc.).
• JSON-based configuration files.
• Hardcoded command-line flags and argparse logic.

Breaking Changes

• Backward compatibility is broken with all 1.x versions.
• You must migrate to the new API and TOML-based configuration system.

ClinicaDL v1.6.1

ClinicaDL 1.6.1

Fixed

• Fix sorting particpants sessions with 2 digit,
• Fix interpret —-save_nifti issue,
• Fix BIDS file format for pet,

Changed

• Change 2 digits session label to 3 digits,
• Change black and isort for ruff and codespell,
• Update type hint and docstring,

New

• Add --fsdp option
• Add --valid_longitudinal option to allow the validation on longitudinal data,

ClinicaDL v1.6.0

pyproject release (#530)

ClinicaDL v1.5.1

ClinicaDL 1.5.1

Fixed

• Fix retrocompatibility for new option in maps.json

Changed

• Change MapsManager architecture by adding Callbacks

New

• Add --fully_sharded_data_parallelism option
• Add --emisions_calculator option with codecarbon
• Add the semi-supervised domain adaptation network proposed for the MICCAI DART workshop.

ClinicaDL v1.5.0

ClinicaDL 1.5.0

Fixed

• Fix adapt command

Changed

• SSIM is now executed on GPU when possible

New

• Add new command generate artifacts to generate noise/contract/motion
• Add options for data augmentation
• Add fine-tuning option
• Add --amp for automatic mixed precision
• Add --track_exp option to track your parameters during training with MLflow or WandB

ClinicaDL v1.4.0

Fixed

• Fix --diagnoses and --merged_tsv option bug in clinicadl get-labels
• Fix Pathlib bugs
• Fix get_tsv_paths function bug.
• Fix a bug for which it was impossible to use predict without specifying the splits and selection metrics.

Changed

• Changed default batch size to 8 for clinicadl predict
• Changed VAEs main class

New

• Add --n_proc option in clinicadl generate pipelines for parallelization.
• Add --split to clinicadl predict
• Add new VAE networks.
• Add pytorch function to summarize
• Add --size_reduction and --size_reduction_factor options to clinicadl train, clinicadl predict and clinicadl interpret.
• Add SSIM2D, SSIM3D metrics for VAE.
• Add --save_latent_space option to clinicadl train and clinicadl predict.
• Add clinicadl generate trvial_motion
• Add Data augmentation with torchio

Release ClinicaDL 1.3.1

ClinicaDL 1.3.1

Fixed

• Fix TypeError when running ClinicaDL.
• Fix --extract_json option bug in clinicadl prepare-data.
• Fix clinicadl tsvtools get-labels error finding clinica iotools missing-modalities output.

Changed

• Changed clinicadl tsvtools get-labels output directory.

New

• Add --caps_directory option in clinicadl tsvtools get-labels.

Release ClinicaDL v1.3.0

ClinicaDL 1.3.0

New

• Add new command quality-check pet-linear.
• Add new command generate hypometabolic.
• Add new network architecture: Resnet3D and SqueezeExcitationCNN.
• Add flair-linear modality for prepare-data command.
• Add pytorch profiler.
• Add --save_nifti option for interpretcommand.
• Add --output_dir argument for tsvtools get-labels command

Changed

** Core: **

• Transition from os to pathlib.
• Update data CI.
• Improve maps_manager.
• Change --acq_label option for --tracer.
• Update tutorial.

ClinicaDL 1.2.0

ClinicaDL 1.2.0

Changed

** Core: **

• Add ClinicaDL installation with pipx.
• Improve logging.
• Add method argument to the interpret command to choose between the new Grad-CAM method and the gradient method.
• Change extract command to prepare-data.
• Change output of get-labels, split and kfold commands to one TSV per split instead of one per label.
• Change tsvtool command to tsvtools.
• Change tsvtools getlabels command to tsvtools get-labels and remove the progression column in the TSV output.
• Add new commands: tsvtools get-progression, tsvtools get-metadata, tsvtools prepare-experiment (split + kfold), and tsvtools adapt.
• Update data CI.
• Add a new model for quality check

Fixed

• Fix quality-check t1-linear