Moonshot’s Kimi K2.5 is 'open,' 595GB, and built for agent swarms — Reddit wants a smaller one

Moonshot AI’s Kimi K2.5 Reddit AMA revealed why the powerful open-weight model is hard to run, plus new details on agent ...
IEEE

MCDiff: A Multilevel Conditional Diffusion Model for PolSAR Image Classification

Abstract: With the swift advancement of deep learning, significant strides have been made in polarimetric synthetic aperture radar (PolSAR) image classification, particularly with the advent of ...
GitHub

DAViD: Modeling Dynamic Affordance of 3D Objects Using Pre-trained Video Diffusion Models (ICCV 2025)

This is the official code for the paper "DAViD: Modeling Dynamic Affordance of 3D Objects Using Pre-trained Video Diffusion Models". Otherwise, you can use open-source image-to-video models such as ...
IEEE

Modeling Spectral LED Degradation Using an Unsupervised Machine Learning Approach

Abstract: The modeling of spectral characteristics of light-emitting diodes (LED) has been addressed in various studies. We extend the current state of knowledge by modeling the spectral ...
Tobacco Control

Cigarette sharing and its association with smoking cessation in China, 2022: findings from the China national adult tobacco survey

Background Cigarette sharing has not been assessed at the provincial level or analysed the degree of its association with smoking cessation in China. Such analysis would provide critical intelligence ...