Back to search
This repository provides a 3D implementation of DINOv2 for self-supervised pretraining on volumetric (3D) medical images using Lightly, MONAI, and Pytorch Lightning!
Stars
52
Forks
6
Watchers
52
Open Issues
5
Overall repository health assessment
No package.json found
This might not be a Node.js project
Merge pull request #16 from AIM-Harvard/copilot/update-docs-with-latest-improvements
87e5e30View on GitHubRemove all DINOv3 and dinotxt/multimodal references from README
deab4f3View on GitHubAddress review feedback: revert What-you-can-do section (add debug bullet only), remove prediction and multimodal sections, expand nnUNet export with nnssl adaptation plan details
912f845View on GitHubUpdate README with comprehensive documentation for all features
efb2e5eView on GitHubMerge pull request #15 from AIM-Harvard/copilot/add-logger-for-gram-matrix
be2fb19View on GitHubAdd fp16-safe projection head and use LR zeroing for last layer warmup
5e9e071View on GitHubMerge pull request #14 from AIM-Harvard/copilot/add-logger-for-gram-matrix
1fff4d4View on GitHubUpdate README with collapse monitoring callbacks, remove callbacks/README
25e2134View on GitHubMerge pull request #13 from AIM-Harvard/copilot/add-logger-for-gram-matrix
8307337View on GitHubMerge pull request #12 from AIM-Harvard/copilot/update-readme-warning
ce5ab35View on GitHub