qyllan

yes, this would be an AMAZING feature to have! it would save me at least 20% of setup time.

something like: ```python wandb.log({"validation_loss": val_loss, "training_loss": train_loss}, same_plot=True) ``` this would automatically hide the `validation_loss` and `training_loss` metrics individually, and on the UI it would show them on a single...

yeah that would be cool if you could put it in a feature request thanks for the response @nate-wandb

I am also having this problem. Have there been any advancements to address this? What are the environment specs for people who have it working? I am getting this issue...

I figured out how to solve it. You want to use PyTorch 1.6.0+ (I also used python3.6) Hope this helps!

yeah, that was the conclusion i came to as well. it's significantly faster than torch's einops still.

> Hey, there is no support for interacting with the world files in Python currently, and atm not planned. It would be a nice feature to have, though, if someone...

quick fix: download 7zip and use the `7zz x files.zip` command

so, #1338 implemented the image encryption algorithm already and it was applied in the context of facebook already, right? as i understand it, my job is now to simply use...

:-)