This code corresponds to the preprint "Reducing the cost of posterior sampling in linear inverse problems via task-dependent score learning".

The following command installs some of the most important python packages

pip install -r requirements.txt

This project includes code from the GitHub repository score_sde_pytorch, which is licensed under the Apache License 2.0.

The following files are included from the score_sde_pytorch repository:

• https://github.com/yang-song/score_sde_pytorch/tree/main/models
• https://github.com/yang-song/score_sde_pytorch/blob/main/utils.py
• https://github.com/yang-song/score_sde_pytorch/blob/main/sde_lib.py
• https://github.com/yang-song/score_sde_pytorch/blob/main/losses.py

score_sde_pytorch repository: https://github.com/yang-song/score_sde_pytorch

To obtain a copy of the necessary files, run

python setup.py

The configs in the folder custom_configs are adaptations of the configs provided in the score_sde_pytorch repository. The original can be found at

• https://github.com/yang-song/score_sde/tree/main/configs

The code from the score_sde_pytorch repository is licensed under the Apache License, Version 2.0. You may obtain a copy of the License at http:https://www.apache.org/licenses/LICENSE-2.0.

The code that performs the training of a score function approximation is in the file training.ipynb.

For generating posterior samples as described in our paper, you can access the code in two separate notebooks: sampling_deblurr.ipynb for deblurring tasks and sampling_ct.ipynb for CT-imaging tasks. These notebooks contain the necessary code to generate posterior samples based on our approach.

Checkpoints of the pretrained unconditional model and the task-dependent models for deblurring and ct-imaging can be found in this Google drive. Download the checkpoints and place them in a folder named checkpoints.

This project is dual-licensed under the MIT License and the Apache License 2.0.

• The code authored for this project is licensed under the MIT License.
• Portions of this codebase are derived from the score_sde_pytorch repository, which is licensed under the Apache License 2.0.

Please see the LICENSE and LICENSE_APACHE files for details.