Welcome to equiformer-pytorch Discussions!

[lucidrains]

👋 Welcome!

We’re using Discussions as a place to connect with other members of our community. We hope that you:

• Ask questions you’re wondering about.
• Share ideas.
• Engage with other community members.
• Welcome others and are open-minded. Remember that this is a community we
  build together 💪.

To get started, comment below with an introduction of yourself and tell us about what you do with this community.

[abdalazizrashid]

Any ideas how to use this model for conditional generation of crystals?

[lucidrains]

i would go the score matching / ddpm route https://arxiv.org/abs/2205.15019

[lucidrains]

start with invariance, as it is a lot cheaper than equivariance https://github.com/lucidrains/invariant-point-attention

[zhuyu-cs]

Hi, is it possiable for predicting the adjacent matrix？Since in molecule-building，the connection information is still important.

[lucidrains]

@zhuyu-cs i'd be doing your research for you, but the route i would take is to extract all the type 1 attention maps, concat them, and do a small MLP head for predicting adjacency

[radiradev]

How does this transformer compare with en_transformer in terms of performance and use cases?