TACO: Ternary Absolute-max Column-wise One-sparse Optimizer for LLM Fine-Tuning

arXiv:2610.02199 · cs.LG, math.OC · Submitted 2026-10-01 · Read on arXiv

cs.LG, math.OC

Submitted: 2026-10-01

Updated: 2026-10-06

Code: https://github.com/Jichao2357/TACO

Terminology

• Fine-Tuning — 1× in this paper · explained in 5 episodes

Sources

• Old Optimizer, New Norm: An Anthology
• signSGD with Majority Vote is Communication Efficient And Fault Tolerant
• 8-bit Optimizers via Block-wise Quantization
• The Llama 3 Herd of Models
• LoRA: Low-Rank Adaptation of Large Language Models
• FP8 Formats for Deep Learning
• SignMuon: Communication-Efficient Distributed Muon Optimization
• Can Muon Fine-tune Adam-Pretrained Models?
• Practical Efficiency of Muon for Pretraining
• Qwen3 Technical Report
• OPT: Open Pre-trained Transformer 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: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks