TAEHV is a Tiny AutoEncoder for Hunyuan Video (and other similar video models). TAEHV can encode and decode latents into videos more cheaply (in time & memory) than the full-size video VAEs, at the ...
Abstract: Hyperspectral unmixing (HU) aims to decompose mixed pixels into their constituent spectral signatures and estimate their corresponding fractional abundances. Recently, the nonlinear spectral ...
Abstract: This article presents a deep autoencoder-based methodology for unsupervised anomaly detection in centrifugal pumps under limited failure data conditions, focusing on real-world applications ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results