Pith. sign in

REVIEW 6 cited by

NOLA: Compressing LoRA using Linear Combination of Random Basis

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.02556 v2 pith:GP6STP4A submitted 2023-10-04 cs.CL cs.CV

NOLA: Compressing LoRA using Linear Combination of Random Basis

classification cs.CL cs.CV
keywords loraranknolalinearmodelsnumberparametersadaptation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Fine-tuning Large Language Models (LLMs) and storing them for each downstream task or domain is impractical because of the massive model size (e.g., 350GB in GPT-3). Current literature, such as LoRA, showcases the potential of low-rank modifications to the original weights of an LLM, enabling efficient adaptation and storage for task-specific models. These methods can reduce the number of parameters needed to fine-tune an LLM by several orders of magnitude. Yet, these methods face two primary limitations: (1) the parameter count is lower-bounded by the rank one decomposition, and (2) the extent of reduction is heavily influenced by both the model architecture and the chosen rank. We introduce NOLA, which overcomes the rank one lower bound present in LoRA. It achieves this by re-parameterizing the low-rank matrices in LoRA using linear combinations of randomly generated matrices (basis) and optimizing the linear mixture coefficients only. This approach allows us to decouple the number of trainable parameters from both the choice of rank and the network architecture. We present adaptation results using GPT-2, LLaMA-2, and ViT in natural language and computer vision tasks. NOLA performs as well as LoRA models with much fewer number of parameters compared to LoRA with rank one, the best compression LoRA can archive. Particularly, on LLaMA-2 70B, our method is almost 20 times more compact than the most compressed LoRA without degradation in accuracy. Our code is available here: https://github.com/UCDvision/NOLA

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. DiV-INR: Extreme Low-Bitrate Diffusion Video Compression with INR Conditioning

    eess.IV 2026-04 unverdicted novelty 7.0

    DiV-INR integrates implicit neural representations as conditioning signals for diffusion models to achieve better perceptual quality than HEVC, VVC, and prior neural codecs at extremely low bitrates under 0.05 bpp.

  2. Chebyshev Manifold Adaptation

    cs.LG 2026-07 conditional novelty 6.0

    ChebyMA uses Chebyshev polynomial surfaces to parameterize weight updates, claiming a better parameter-accuracy trade-off than LoRA, TLoRA, and StelLA on CIFAR and text classification.

  3. DECA: Decentralizing Block-Wise Adam for Efficient LLM Full-Parameter Fine-Tuning on Non-IID Data

    cs.LG 2026-06 unverdicted novelty 5.0

    DECA partitions LLM parameters into blocks for sequential block-wise Adam optimization in decentralized non-IID settings to support efficient full-parameter fine-tuning.

  4. Mask the Target: A Plug-and-Play Regularizer Against LoRA Forgetting

    cs.CL 2026-05 unverdicted novelty 5.0

    A plug-and-play KL regularizer that masks the target token and renormalizes probabilities to improve the learning-forgetting trade-off in LoRA adaptation of LLMs.

  5. Multi-Modal Learning meets Genetic Programming: Analyzing Alignment in Latent Space Optimization

    cs.NE 2026-04 unverdicted novelty 5.0

    SNIP's symbolic-numeric alignment stays coarse and does not improve during optimization, so multi-modal LSO does not yet deliver effective bi-modal search for symbolic regression.

  6. Multi-Modal Learning meets Genetic Programming: Analyzing Alignment in Latent Space Optimization

    cs.NE 2026-04 unverdicted novelty 4.0

    Experiments reveal that cross-modal alignment in SNIP does not improve with increasing fitness and is too coarse for effective symbolic search in latent space optimization for symbolic regression.