🛰️ Daily AI Frontier
‹ back to 2026-08-24

ProxyFormer: A Dual-Stream Proxy Architecture for Ultra-Long Context and High-Resolution Generation

Research Efficiency & Systems

Ranking

Overall 78
Content 95
Popularity 37

Observed public metrics from 1 member.

Representative image for ProxyFormer: A Dual-Stream Proxy Architecture for Ultra-Long Context and High-Resolution Generation

Merged summary

TL;DR - ProxyFormer uses persistent local features alongside compressed proxy tokens to make global attention and KV caching substantially cheaper. It enables million-token retrieval and much longer training contexts on limited GPU memory while also showing preliminary applicability to high-resolution image generation.

  • Each layer compresses local features into proxy states, performs global interactions in proxy space, then injects the contextualized proxies back into the persistent local stream.
  • Factorized multi-level compression, dynamic layer-wise compression ratios, asymmetric embeddings, and proxy-only KV caching further reduce resource requirements.
  • On a 16GB GPU, a compression ratio of 64 increases the reported trainable context from roughly 20K to 0.7M tokens.
  • A 64K-window model retains 92%–95% multi-needle retrieval accuracy at 1,048,576 tokens; preliminary flow-matching experiments cover pixel and latent image generation.

Sources (1)

ProxyFormer: A Dual-Stream Proxy Architecture for Ultra-Long Context and High-Resolution Generation

arXiv cs.LG Zhongpan Tang 2026-08-24 arXiv:2608.23463
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-14 14:17:47.315108 UTC

TL;DR - ProxyFormer uses persistent local features alongside compressed proxy tokens to make global attention and KV caching substantially cheaper. It enables million-token retrieval and much longer training contexts on limited GPU memory while also showing preliminary applicability to high-resolution image generation.

  • Each layer compresses local features into proxy states, performs global interactions in proxy space, then injects the contextualized proxies back into the persistent local stream.
  • Factorized multi-level compression, dynamic layer-wise compression ratios, asymmetric embeddings, and proxy-only KV caching further reduce resource requirements.
  • On a 16GB GPU, a compression ratio of 64 increases the reported trainable context from roughly 20K to 0.7M tokens.
  • A 64K-window model retains 92%–95% multi-needle retrieval accuracy at 1,048,576 tokens; preliminary flow-matching experiments cover pixel and latent image generation.
item →