AIPython★ 746⑂ 52 forksMIT
fudan-generative-vision/Remory
Residual Memory Context Compaction
Remory trains a residual memory module to compact long contexts into shorter representations for transformer models.
// KEY FEATURES
- Learns residual memory to compress context while retaining key information.
- Outputs compact token sequences suitable for transformer input layers.
- Enables longer-context processing with reduced memory and compute overhead.
New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram