-
hi authors, thanks for the great work!
i just wonder if LR=1e-3 for mup is optimal value from small-scale proxy model
and how dropout is critical for multi-epoch training.
for the latter, i guess y…
-
`mup` is currently the only automated way to deploy `patients`. `mup` is coupled to Meteor and we want to get rid of Meteor. A solution based on `docker-compose` should be sufficient.
-
Hello @zodern, I recently migrated my project to Meteor 3.0 and encountered an issue. I found out where the problem is coming from, but what can we do as a solution?
I updated my application sequen…
-
# 🚀 Feature request
This request is to open up a discussion on 1) whether it makes sense to implement [Maximal Update Parametrization (abbreviated muP)](http://arxiv.org/abs/2203.03466) in Hugg…
-
https://github.com/cloneofsimo/minRF/blob/261859e8b89a4cf5ab7eb35b4a4ffd8037c35ea1/advanced/mmdit.py#L161
https://github.com/cloneofsimo/minRF/blob/72feb0c87d435e9f9d220f34f348ed66c0b6ccec/advanced/m…
-
Hi, first of all thanks for this wonderful package.
However, I am facing issues while using this with MUP. I get the following error during deployment (sorry if this is an issue with MUP). but it wor…
ytay2 updated
8 years ago
-
Hi there
is someone still using this package?
nowadays with mup, meteor 2.x etc?
Would be interesting since in principle this package solves the scaling problem elegantly :)
KR
-
Is there anybody success deploy with [mup](https://github.com/arunoda/meteor-up)? I always got error :
```
-----------------------------------STDERR----------------------------------- …
-
Is mup compatible with torch.compile() in Pytorch 2? If yes, what is the correct usage (e.g. should we apply mup before compile or after)?
-
How do I get this command to work? I think my version is out of date or something. I have tried to update this program but I'm not sure if I'm up to date, or what..
If I type mup , this is what I g…