You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have to simplify the output directory situation, we should override in some way the hydra-core file structure, ideally, our custom run name/tags should be first, and differentiation based by date/hour should be deeper in the file structure, making it easier to read from tensorboard.
As a second step, we should add a more easy-to-use parameter to override/replace/append to the custom/aux path to make it much easier to have custom names for our runs, moving "aux_save_dir" at the start of save_dir, could be the simplest way (and rename it)
So we can run things like: "python main.py tag=increased_lr_same_model" and to have this be easily readable from tensorboard.
The text was updated successfully, but these errors were encountered:
We have to simplify the output directory situation, we should override in some way the hydra-core file structure, ideally, our custom run name/tags should be first, and differentiation based by date/hour should be deeper in the file structure, making it easier to read from tensorboard.
As a second step, we should add a more easy-to-use parameter to override/replace/append to the custom/aux path to make it much easier to have custom names for our runs, moving "aux_save_dir" at the start of save_dir, could be the simplest way (and rename it)
So we can run things like: "python main.py tag=increased_lr_same_model" and to have this be easily readable from tensorboard.
The text was updated successfully, but these errors were encountered: