Re:Stable Diffusion in pure C/C++
by waynecochran on Saturday August 19, 2023 @01:53PM
(#37191320)
Here is a tiny excerpt try to get dvc to work just so I could get the training weights for deployment ... remember I don't develop much w Python...
Finally I just have my colleague manually copy the weights. This kind of thing went for hours.
[ Parent
| on HN ]
Researchers are notorious for writing bad code
What even is dvc
Re:Stable Diffusion in pure C/C++
by waynecochran on Saturday August 19, 2023 @02:26PM
(#37191622)
I think dvc is like git for large binary files. You need someway to manage your NN weights -- what are other methods?
git lfs is what everyone is using, HF in particular
Hey, DVC maintainer here.
Thanks for giving DVC a try!
With snap, you need to use `--classic` flag, as noted in https://dvc.org/doc/install/linux#install-with-snap Unfortunately that's just how snap works for us there :(
Regarding the pull error, it simply looks like you don't have some credentials set up. See https://dvc.org/doc/user-guide/data-management/remote-storag... Still, the error could be better, so that's on us.
Feel free to ping us in discord (see invite link in https://dvc.org/support). I'm @ruslan there. We'll be happy to help.
Thanks, Ruslan
Re:Stable Diffusion in pure C/C++
by waynecochran on Sunday August 20, 2023 @09:24AM
(#37198908)
Thanks… i know my colleague uses it a lot. I generally use his models and don’t do much ML development yet. At some point I need to properly learn all of this. It seems ML tools are only for developers not for those who simply want to deploy and use the resulting NN.