attention-residuals

Tag

Cards List
#attention-residuals

SANA-Video 2.0: Hybrid Linear Attention with Attention Residuals for Efficient Video Generation

Hugging Face Daily Papers · 2026-07-23 Cached

SANA-Video 2.0 introduces a hybrid linear-softmax attention mechanism for video diffusion transformers, achieving high-quality video generation up to 720p on a single GPU with significantly reduced latency compared to full-softmax models, while maintaining competitive VBench scores.

0 favorites 0 likes
#attention-residuals

Multi-Head Attention Residuals

Hugging Face Daily Papers · 2026-07-22 Cached

Introduces Multi-Head Attention Residuals (MHAR), which reshapes the routing query into per-subspace heads so each feature subspace reads depth history via its own softmax. Trained from scratch on a Nemotron-based corpus, MHAR consistently improves validation loss over standard Transformers from 100M to 1B scale and also boosts mid-training downstream accuracy.

0 favorites 0 likes
#attention-residuals

Low-Rank Attention Residuals

arXiv cs.LG · 2026-07-14 Cached

This paper introduces Low-Rank Attention Residuals (LR-AttnRes) for LLMs, which decouple routing from representation by using low-dimensional keys for depth-wise attention, improving performance while reducing FLOPs.

0 favorites 0 likes
#attention-residuals

@gurtej__gill_: The Kimi Team wrote a really clever paper back in March that fixes a fundamental flaw we have sort of just accepted in …

X AI KOLs Timeline · 2026-07-03 Cached

The Kimi Team's paper 'Attention Residuals' (AttnRes) replaces uniform residual connections in Transformers with softmax attention over depth, allowing each layer to dynamically select earlier representations. Pre-trained on 1.4 trillion tokens with a 48B parameter model, it stabilizes hidden states and significantly improves reasoning tasks.

0 favorites 0 likes
#attention-residuals

Delta Attention Residuals

Hugging Face Daily Papers · 2026-05-13 Cached

Delta Attention Residuals improve layer-wise routing in transformer models by attending to feature changes (deltas) rather than cumulative hidden states, achieving 1.7–8.2% validation perplexity gains across scales from 220M to 7.6B parameters.

0 favorites 0 likes
← Back to home

Submit Feedback