chaitjo commited on
Commit
80e841a
·
verified ·
1 Parent(s): 84aff32

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +42 -3
README.md CHANGED
@@ -1,3 +1,42 @@
1
- ---
2
- license: mit
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ datasets:
4
+ - chaitjo/QM9_ADiT
5
+ - chaitjo/MP20_ADiT
6
+ - chaitjo/QMOF150_ADiT
7
+ tags:
8
+ - chemistry
9
+ - materials
10
+ - molecules
11
+ - crystals
12
+ - diffusion
13
+ - transformer
14
+ - latent-diffusion
15
+ - all-atom
16
+ ---
17
+
18
+ # All-atom Diffusion Transformers
19
+
20
+ [![arXiv](https://img.shields.io/badge/PDF-arXiv-blue)](https://www.arxiv.org/abs/2503.03965)
21
+ [![Code](https://img.shields.io/badge/Code-GitHub-blue)](https://github.com/facebookresearch/all-atom-diffusion-transformer/)
22
+ [![X](https://img.shields.io/badge/X_thread-@chaitjo-blue)](https://x.com/chaitjo/status/1899114667219304525)
23
+ [![Slides](https://img.shields.io/badge/Slides-chaitjo.com-blue)](https://www.chaitjo.com/publication/joshi-2025-allatom/All_Atom_Diffusion_Transformers_Slides.pdf)
24
+
25
+ Independent reproduction of the paper [*"All-atom Diffusion Transformers: Unified generative modelling of molecules and materials"*](https://www.arxiv.org/abs/2503.03965), by [Chaitanya K. Joshi](https://www.chaitjo.com/), [Xiang Fu](https://xiangfu.co/), [Yi-Lun Liao](https://www.linkedin.com/in/yilunliao), [Vahe Gharakhanyan](https://gvahe.github.io/), [Benjamin Kurt Miller](https://www.mathben.com/), [Anuroop Sriram*](https://anuroopsriram.com/), and [Zachary W. Ulissi*](https://zulissi.github.io/) from FAIR Chemistry at Meta (* Joint last author).
26
+
27
+ All-atom Diffusion Transformers (ADiTs) jointly generate both periodic materials and non-periodic molecular systems using a unified latent diffusion framework:
28
+ - An autoencoder maps a unified, all-atom representations of molecules and materials to a shared latent embedding space; and
29
+ - A diffusion model is trained to generate new latent embeddings that the autoencoder can decode to sample new molecules or materials.
30
+
31
+ ## Citation
32
+
33
+ ArXiv link: [*All-atom Diffusion Transformers: Unified generative modelling of molecules and materials*](https://www.arxiv.org/abs/2503.03965)
34
+
35
+ ```
36
+ @article{joshi2025allatom,
37
+ title={All-atom Diffusion Transformers: Unified generative modelling of molecules and materials},
38
+ author={Chaitanya K. Joshi and Xiang Fu and Yi-Lun Liao and Vahe Gharakhanyan and Benjamin Kurt Miller and Anuroop Sriram and Zachary W. Ulissi},
39
+ journal={arXiv preprint},
40
+ year={2025},
41
+ }
42
+ ```