Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/nicholas-miklaucic/avid
https://github.com/nicholas-miklaucic/avid
Last synced: 13 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/nicholas-miklaucic/avid
- Owner: nicholas-miklaucic
- Created: 2024-02-28T22:12:04.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2024-04-12T04:33:21.000Z (7 months ago)
- Last Synced: 2024-04-12T10:00:06.048Z (7 months ago)
- Language: Jupyter Notebook
- Size: 12.5 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AVID
## Acronym Very Incomplete: Diffusion?
Work on applying diffusion and vision transformers to materials. Right now, I'm
just doing cubic perovskites. Cool stuff TBD.## Installation
[Install JAX](https://jax.readthedocs.io/en/latest/installation.html), and then
install the other necessary libraries:```bash
pip install toml pyrallis rho-plus equinox diffrax optax plotext rich einops pandas plotly matplotlib seaborn pymatgen tfp-nightly[jax] textual textual-dev textual-plotext flax orbax beartype humanize clu pyarrow
```TODO:
- [ ] Perhaps configure DiLED in addition to normal diffusion and encoder losses
- [ ] How to train the encoder? As part of the normal process or separately?
- [ ] EMA
- [ ] Refactor diffusion trainer into TrainingRun so we can get dashboards again