Test-time Regression Improves Sequence Compression
TTR-Compressor uses end-to-end learned regression for sequence compression, outperforming pooling and projections in language and multimodal models.
2 stories, newest first · one card per story
TTR-Compressor uses end-to-end learned regression for sequence compression, outperforming pooling and projections in language and multimodal models.
On visual token merging it matches an MLP merger while using 55.6% fewer visual tokens.
Today 1.5K readers · 2.4K views · 757 reads · all time 189K views