Tetra: Serving Leech-Lattice Quantized LLMs at 2.7 Bits per Parameter
cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
Code: https://github.com/pjmalandrino/llvq
Sources
- GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
- QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMs
- QuIP: 2-Bit Quantization of Large Language Models With Guarantees
- Training Verifiers to Solve Math Word Problems
- Extreme Compression of Large Language Models via Additive Quantization
- GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
- MARLIN: Mixed-Precision Auto-Regressive Parallel Inference on Large Language Models
- Measuring Massive Multitask Language Understanding
- Distilling the Knowledge in a Neural Network
- Efficient Memory Management for Large Language Model Serving with PagedAttention
- AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration
- VPTQ: Extreme Low-bit Vector Post-Training Quantization for Large Language Models
- Qwen3 Technical Report
- QuIP#: Even Better LLM Quantization with Hadamard Incoherence and Lattice Codebooks
- QTIP: Quantization with Trellises and Incoherence Processing
- Leech Lattice Vector Quantization for Efficient LLM Compression
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: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks