Philip Meier

@mthrok I took the liberty to auto-cc you every time I label a PR as `prototype` so I don't have to remember to do it manually. For now this will...

I've added the needed changes for the detection references.

I've run the references for a few iterations with the following parameters to confirm they work: - Classification: ```py [ "--device=cpu", "--batch-size=2", "--epochs=1", "--workers=2", "--mixup-alpha=0.5", "--cutmix-alpha=0.5", "--auto-augment=ra", # "ra", "ta_wide",...

~Detection references are affected by #6528. Do not train before this is merged.~

*This comment only applies in case we decide to reinstate the high-level dispatchers. Leaving this here so it doesn't get lost.* --- Since torchscript is statically compiled we need special...

Regardless of #6584, the transforms will not be JIT scriptable. Thus, cleaning them up is still relevant.

I don't think the proposal was ever to wrap `itertools` inside the pipes, but rather to mirror the functionality and naming. So if the function is indeed stateless, we can...

> Thank god for this Well, you could thank me instead :stuck_out_tongue: Jokes aside, thanks for posting this. It means a lot. I'm glad that this project helped you out....

> My only request would be to add torchvision (and maybe torchaudio) as extra possible installs along with? I held my breath when I added it but it worked ok....

Hey @amitkparekh :wave: > I'm using this package within a Dockerfile to automatically install the correct CUDA version for the system but I'm having issues getting the build process to...