mlr3torch
mlr3torch copied to clipboard
Deep learning framework for the mlr3 ecosystem based on torch
* [ ] pass parameters to dataloader * [ ] Specify how often to do validation (every epoch probably not always needed).
* [ ] clean up references
This should maybe be an issue in `mlr3pipelines`, but I am just writing down my idea here. The problem is that validation does not work when being in a `Graph`,...
Also consider this: https://github.com/mlr-org/mlr3/pull/977/files
* [ ] Change status in the overview page
Some predefined callbacks we might want to support: * [ ] LR scheduler * [ ] Writing files in tensorboard format: https://github.com/mlverse/tfevents * [ ] Experiment Tracking using guildai-r: https://github.com/guildai/guildai-r...
In the weekly run however we want to ensure that the downloads still work