Skip to content

Latest commit

 

History

History
20 lines (20 loc) · 973 Bytes

TODO.md

File metadata and controls

20 lines (20 loc) · 973 Bytes

[X] Download NLB datasets [X] Double-check that masks are properly dealt with 3 subtypes of masking [X] Fix issues with Mac-based training ("Cannot convert a MPS tensor to float64 dtype as the MPS framework doesn't support float64) [X] Install ipython locally, jupyter [X] Add compatibility with NLB dataset [X] Test out prep_nlb.py [X] Add more information about the dataset and the problem into the deck [X] Add more information about me in the deck [X] Install M1 local GPU based training with PyTorch nightlies`` [X] Check that checkpoint loading works as expected [X] Look into Convnext with MLM methodology [ ] Finish presentation [ ] Re-write mask pedagogically [ ] Implement augmentation (e.g. dropout) [ ] Create branches listed in README.md [ ] Implement some of the suggestions listed in README.md [ ] Implement Hydra-based configuration [ ] Try running under SLURM on Mila infra [ ] Implement alternative method with mask [ ] Move tests out of the main branch