nv-generate-mr

To generate synthetic MRI volumes for benchmarking, run the wrapper script with a JSON config file and output directory. It validates prerequisites, shells out to the MR workflow, and produces an auditable JSON payload summarizing results.

Synthetic MR image synthesis is used for research, benchmarking, and governance of MR pipelines across configurations, modalities, and seeds. This wrapper provides engineering validation and structured provenance for safe, repeatable generation.

Yes, you need CUDA and PyTorch to run synthetic MRI generation. The wrapper performs preflight checks for CUDA availability and model inventory, and it requires dependencies like torch and numpy to execute safely.

Yes, the MRI synthesis wrapper supports NIfTI output for medical imaging. It depends on nibabel for handling NIfTI files and generates synthetic MR volumes validated through the configured pipeline.

To prevent expensive synthetic MRI generation runs, the wrapper includes preflight checks with safeguards and cost estimates. It evaluates the configuration and environment before shelling out to the upstream CLI.

To track provenance and configuration for synthetic MRI generation, the wrapper outputs a detailed JSON payload. This includes the upstream commit, invocation details, staged configs, and results for governance.