Abstract:Large language models (LLMs) are notoriously memory-intensive during training, particularly with the popular AdamW optimizer. This memory burden necessitates using more or higher-end GPUs or reducing batch sizes, limiting training scalability and throughput. To address this, various memory-efficient optimizers have been proposed to reduce optimizer memory usage. However, they face critical challenges: (i) reliance on costly SVD operations; (ii) significant performance trade-offs compared to AdamW; and (iii) still substantial optimizer memory overhead to maintain competitive performance. In this work, we identify that AdamW's learning rate adaptation rule can be effectively coarsened as a structured learning rate update. Based on this insight, we propose Approximated Gradient Scaling for Memory-Efficient LLM Optimization (APOLLO), which approximates learning rate scaling using an auxiliary low-rank optimizer state based on pure random projection. This structured learning rate update rule makes APOLLO highly tolerant to further memory reductions while delivering comparable pre-training performance. Even its rank-1 variant, APOLLO-Mini, achieves superior pre-training performance compared to AdamW with SGD-level memory costs. Extensive experiments demonstrate that the APOLLO series performs on-par with or better than AdamW, while achieving greater memory savings by nearly eliminating the optimization states of AdamW. These savings provide significant system-level benefits: (1) Enhanced Throughput: 3x throughput on an 8xA100-80GB setup compared to AdamW by supporting 4x larger batch sizes. (2) Improved Model Scalability: Pre-training LLaMA-13B with naive DDP on A100-80GB GPUs without system-level optimizations. (3) Low-End GPU Friendly Pre-training: Pre-training LLaMA-7B on a single GPU using less than 12 GB of memory with weight quantization.

Training Neural Networks with Low-Precision Model Memory

Low-Memory Neural Network Training: A Technical Report

Memory Efficient Optimizers with 4-Bit States

Memory Efficient Mixed-Precision Optimizers

Efficient Adaptive Optimization via Subset-Norm and Subspace-Momentum: Fast, Memory-Reduced Training with Convergence Guarantees

APOLLO: SGD-like Memory, AdamW-level Performance

ZO-AdaMU Optimizer: Adapting Perturbation by the Momentum and Uncertainty in Zeroth-Order Optimization

Progressive Learning of Low-Precision Networks

Low-Rank Training of Deep Neural Networks for Emerging Memory Technology

Training with Mixed-Precision Floating-Point Assignments

The ZipML Framework for Training Models with End-to-End Low Precision: The Cans, the Cannots, and a Little Bit of Deep Learning

Optimizing for In-memory Deep Learning with Emerging Memory Technology

Collage: Light-Weight Low-Precision Strategy for LLM Training

AdaLomo: Low-memory Optimization with Adaptive Learning Rate

Reducing Fine-Tuning Memory Overhead by Approximate and Memory-Sharing Backpropagation

Speeding up and reducing memory usage for scientific machine learning via mixed precision

Revisiting Zeroth-Order Optimization for Memory-Efficient LLM Fine-Tuning: A Benchmark

Recurrent Neural Networks With Limited Numerical Precision

DNN Memory Footprint Reduction via Post-Training Intra-Layer Multi-Precision Quantization

Global Sparse Momentum SGD for Pruning Very Deep Neural Networks