Highlights
- Arctic Code Vault Contributor
- Pro
Create your own GitHub profile
Sign up for your own profile on GitHub, the best place to host code, manage projects, and build software alongside 50 million developers.
Sign up
Pinned
2,138 contributions in the last year
Activity overview
Contributed to
PyTorchLightning/pytorch-lightning,
PyTorchLightning/pytorch-lightning-bolts,
ultralytics/yolov5
and 5 other
repositories
Contribution activity
September 2020
Created a pull request in CBICA/HistoReg that received 8 comments
add CI - build & run [wip]
+62
−2
•
8
comments
- Added check for apex AMP and unit tests for Horovod + AMP
- [Metrics] Class reduction similar to sklearn
- Get experiment_id from MLFlow only once instead of each training loop.
- stable, dev PyTorch in Dockerfile and conda gh actions
- [Metrics] AUROC error on multilabel + improved testing
- [Metrics] class based embedding similarity + tests
- Fix yaml deprecation
- Fix sample code of LightningModule DataLoaders
- added emb similarity
- Fix IoU score for classes not present in target or pred
- Fix non autoplaying videos in Doc on Safari iOS
- update batch size in LightningModule.datamodule when auto scaling batch size
- Fix ModelCheckpoints name formatting
- Fix: gather_all_tensors cross GPUs in DDP
- [Blocked by refactors]Fix batch_outputs with optimizer frequencies
- Changed LearningRateLogger to LearningRateMonitor
- use fsspec instead of gfile for all IO
- Early stopping doc
- Docs/improved multigpu doc clarity
- to_torchscript method for LightningModule
- Recursive inspection of subclasses init in `from_argparse_args`
- Add torch jit ignore decorator to Lightning properties
- ref: moved argparse code to central class
- distinguished between overfit_batches examples
- ref: move train outside of setup training
Created an issue in ultralytics/yolov5 that received 5 comments
How suspend image mixing
5
comments