Residual Decoder Adapter upgrades existing visual tokenizers for AR image models by adding a paired codebook and pixel-space residual branch, raising OCR accuracy on text rendering benchmarks from ~12-24% to ~36-58%.
arXiv preprint arXiv:2205.03650 (2022)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it