Flux diffusion model implementation using quantized fp8 matmul & remaining layers use faster half precision accumulate, which is ~2x faster on consumer devices.

Unlike other deep learning (DL) models, Transformer has the ability to extract long-range dependency features from hyperspectral image (HSI) data. Masked autoencoder (MAE), which is based on ...

We propose a Crystal Diffusion Variational Autoencoder (CDVAE) that captures the physical inductive bias of material stability. By learning from the data distribution of stable materials, the decoder ...

In this paper, a signal-guided masked autoencoder (S-MAE) based semi-supervised learning framework is proposed for high-precision positioning with limited labeled channel impulse response (CIR) ...