---
license: mit
---

# Can We Scale Transformers to Predict Parameters of Diverse ImageNet Models?
[Boris Knyazev](http://bknyaz.github.io/), [Doha Hwang](https://mila.quebec/en/person/doha-hwang/), [Simon Lacoste-Julien](http://www.iro.umontreal.ca/~slacoste/)

https://arxiv.org/abs/2303.04143

See https://github.com/SamsungSAILMontreal/ghn3 for the examples on how to use our GHN-3 model.