Appa is a large weather model (1.5B) composed of an autoencoder and a latent diffusion model. After unsupervised training, it can assimilate any kind of non-linear observations to generate plausible trajectories, in tasks such as reanalysis or forecasting. We demonstrate its strong and promising results and flexibility.
NeuralMPM can emulate point-cloud fluid systems that include multiple materials, each with their specific properties, at over 1000 FPS, vs 15 for the original simulator. It only needs trajectories of positions and velocities to learn, bypassing the tricky tuning of a simulator.
Considering the lack of existing public solutions and data for dynamic reconstruction of soccer scenes, we explore the use of the recent neural radiance fields (NeRFs) to tackle this challenging and specific task.
We merge several ideas from the neural style transfer (NST) literature to build consistent video NST. Notably, we use a U-Net, noise injection, temporal losses, self-supervised learning through style and content losses through embeddings into a pre-trained large CNN (VGG-19). This project was done in the Deep Learning class at ULiège. The project was graded 19/20 and best of the year.
Ultra Cosmetics is my largest standalone project, started at 15 years old. It allows servers to be monetized without breaking Mojang's EULA, all for free with a large open-source community.
The project was taken over by Datatags, thanks to him!
Design stolen from Jon Barron's website. Source code