DMRI

DMRI provides diffusion MRI simulators and tools for model selection and parameter inference. The project is pre-release software, and its API and CLI may change.

Early development. Expect rough edges and breaking changes in public API.

Install

DMRI requires Python 3.11 or newer. From a clone of the repository:

uv venv -p 3.11
source .venv/bin/activate
uv pip install -e .

Supported Linux installations include JAX's bundled CUDA 13 runtime by default. See the installation and quickstart for driver requirements and the required input layout.

Predict

dmri predict FOLDER --model msb3s_2_4_6_128 --quality balanced --non-interactive

FOLDER must contain data.nii.gz, nodif_brain_mask.nii.gz, bvals, and bvecs. Prediction writes results below FOLDER/dmri_output/.

• Prediction documentation
• Training documentation
• Evaluation documentation

Development checks

Install development dependencies with uv pip install -e '.[dev]', then run:

pytest
ruff check .
ruff format --check .

Documentation: https://www.mackelab.org/dmri/

License: MIT