TTB: Test-time Baking for Efficient Rendering of Decoder-only View Synthesis Models
TTB shifts cross-view attention into fast-weight MLPs, delivering constant-complexity novel-view rendering while retaining much of decoder-only synthesis quality.
ls ./papers | grep "#computer vision"
Trending AI research papers from Hugging Face, summarized in plain language: what each paper does, why it matters and where to read more.
2 papers · #computer vision
TTB shifts cross-view attention into fast-weight MLPs, delivering constant-complexity novel-view rendering while retaining much of decoder-only synthesis quality.
MatAnyone-E introduces sparse memory propagation and a lightweight decoder to enable efficient, high-quality 4K video matting with 7.4× speedup and 13× less GPU memory.