mae

Tag

Cards List
#mae

Image Tokenizers Should Be Trained Between Images

Hugging Face Daily Papers ↗ · 2026-10-12 Cached

提出InterpTok,通过在训练样本之间的代码混合空间中训练解码器来改进图像tokenizer的插值与生成质量,在MAETok基础上将插值质量提升三倍以上,并将扩散transformer的生成FID从11.61降至8.48(无引导)。

0 favorites 0 likes
#mae

What masking geometry works best for EEG foundation models?

Hugging Face Daily Papers ↗ · 2026-09-27 Cached

This paper systematically evaluates masking strategies for EEG foundation models across MAE and JEPA frameworks, identifying optimal configurations and a novel failure mode, achieving comparable performance to REVE with reduced compute.

0 favorites 0 likes
← Back to home

Submit Feedback