FlashLoop: Fast and Memory-Efficient Looped Transformers via Lazy Updates
cs.LG
Submitted: 2026-09-24
Updated: 2026-09-24
Code: https://github.com/Superone77/FlashLoop
Project page: https://superone77.github.io/FlashLoop
Terminology
Sources
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
- Training Verifiers to Solve Math Word Problems
- Universal Transformers
- The Llama 3 Herd of Models
- Let's Verify Step by Step
- KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
- Looped Latent Attention: Cross-Loop KV Compression for Looped Transformers
- SparQ Attention: Bandwidth-Efficient LLM Inference
- WinoGrande: An Adversarial Winograd Schema Challenge at Scale
- Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference
- Gemma 3 Technical Report
- Memory-Efficient Looped Transformer: Decoupling Compute from Memory in Looped Language Models
- Chain-of-Experts: Unlocking the Communication Power of Mixture-of-Experts Models
- Qwen3 Technical Report
- AlphaQ: Calibration-Free Bit Allocation for Mixture-of-Experts Quantization
- Scaling Latent Reasoning via Looped Language Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks