Exploring parameter-efficient fine-tuning (PEFT) of billion-parameter vision models with QLoRA and DoRA: insights into generalization for limited-data image classification under a 98:1 test-to-train regime

• URL: http://arxiv.org/abs/2603.17782v1
• Date: Wed, 18 Mar 2026 14:46:07 GMT
• Title: Exploring parameter-efficient fine-tuning (PEFT) of billion-parameter vision models with QLoRA and DoRA: insights into generalization for limited-data image classification under a 98:1 test-to-train regime
• Abstract summary: This study systematically compared three approaches: training from scratch (ResNet-18, ViT-Small), frozen feature extraction, and parameter-efficient fine-tuning (PEFT) of the DINOv3 foundation model (6.7 billion parameters)<n>We evaluated QLoRA and DoRA across multiple configurations varying rank (8, 16, 64) and target modules (q_proj versus all-linear layers)<n>Results demonstrated that PEFT substantially outperformed alternatives, where the best QLoRA configuration (all-linear layers and rank=64) achieved 83.16% test accuracy with only 2.72% parameters (183.0M) in 5.8 hours
• Score: 0.5765148824909591
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Automated behavior classification is essential for precision livestock farming but faces challenges of high computational costs and limited labeled data. This study systematically compared three approaches: training from scratch (ResNet-18, ViT-Small), frozen feature extraction, and parameter-efficient fine-tuning (PEFT) of the DINOv3 foundation model (6.7 billion parameters). We evaluated QLoRA and DoRA across multiple configurations varying rank (8, 16, 64) and target modules (q_proj versus all-linear layers). With 2,160 verified training images, we assessed generalization of our model on 211,800 test samples, which is essentially a 98:1 test-to-train ratio. Results demonstrated that PEFT substantially outperformed alternatives, where the best QLoRA configuration (all-linear layers and rank=64) achieved 83.16% test accuracy with only 2.72% parameters (183.0M) in 5.8 hours, compared to 72.87% for ResNet-18 (16.8 hours), 61.91% for ViT-Small (18.7 hours), and 76.56% for frozen DINOv3 (17.5 hours). DoRA achieved comparable accuracy (83.14%) but with longer training time (11.0 hours). Notably, increasing adapter capacity consistently improved generalization while simultaneously not causing overfitting: reducing rank from 16 to 8 decreased test accuracy from 78.38% to 77.17%, while expanding from q_proj-only to all-linear layers with rank=64 improved accuracy from 78.38% to 83.16%. This suggests underfitting, instead of overfitting, is the primary challenge when adapting foundation models to agricultural imagery. Our findings provide guidelines for deploying billion-parameter vision models with PEFT in agricultural livestock applications.

Related papers

• Interpretable machine learning for predicting splitting strength of asphalt concrete: insights from SHAP analysis [2.5274533414341493]
  This paper presents an interpretable machine-learning framework for predicting the splitting strength (ST) of asphalt concrete.<n>Six machine-learning models, namely TabPFN, ANN, SVR, RF, XGBoost, and LightGBM, were developed and compared.<n>Results show that TabPFN delivered the best overall performance on the testing set, with the lowest RMSE of 0.28, MAE of 0.21, MAPE of 18.01%, MAD of 0.14, the highest R2 of 0.88, and the highest composite score of 0.91.
  arXiv  Detail & Related papers  (2026-08-02T03:05:54Z)
• StableTTA: Training-Free Test-Time Adaptation that Improves Model Accuracy on ImageNet1K to 96% [4.467218412104476]
  Ensemble methods are widely used to improve predictive performance, but their effectiveness often comes at the cost of increased memory usage and computational complexity.<n>We propose test-time adaptation (StableTTA), a training-free method employs novel image and logit processing.
  arXiv  Detail & Related papers  (2026-04-06T09:21:48Z)
• Time-Series at the Edge: Tiny Separable CNNs for Wearable Gait Detection and Optimal Sensor Placement [3.7765281299298015]
  We study on-device time-series analysis for gait detection in Parkinson's disease (PD) from short windows of triaxial acceleration, targeting resource-latency wearables and edge nodes.<n>We compare magnitude thresholding to three 1D CNNs for time-series analysis: a literature baseline (separable convolutions) and two ultra-light models - one purely separable and one with residual connections.
  arXiv  Detail & Related papers  (2025-11-29T08:52:41Z)
• MapReduce LoRA: Advancing the Pareto Front in Multi-Preference Optimization for Generative Models [52.32146943039743]
  Reinforcement learning from human feedback has advanced alignment of generative models to human aesthetic and perceptual preferences.<n>We introduce two complementary methods: MapReduce LoRA and Reward-aware Token Embedding (RaTE)<n>Our framework sets a new state-of-the-art multi-preference alignment recipe across modalities.
  arXiv  Detail & Related papers  (2025-11-25T18:49:21Z)
• Beyond Correctness: Evaluating Subjective Writing Preferences Across Cultures [87.75098311090642]
  Current preference learning methods achieve high accuracy on standard benchmarks but exhibit significant performance degradation when objective quality signals are removed.<n>We introduce WritingPreferenceBench, a dataset of 1,800 human-annotated preference pairs (1,200 English, 600 Chinese) across 8 creative writing genres.
  arXiv  Detail & Related papers  (2025-10-16T12:23:13Z)
• Exploring the Hierarchical Reasoning Model for Small Natural-Image Classification Without Augmentation [51.56484100374058]
  It is evaluated on MNIST, CIFAR-10, and CIFAR-100 under a deliberately raw regime.<n>It is concluded that, for small-resolution image classification without augmentation, HRM is not competitive with even simple convolutional architectures.
  arXiv  Detail & Related papers  (2025-10-04T01:22:41Z)
• A Comparative Benchmark of Real-time Detectors for Blueberry Detection towards Precision Orchard Management [2.667064587590596]
  This study presents a novel comparative benchmark analysis of advanced real-time object detectors.<n>This dataset comprises 661 canopy images collected with smartphones during the 2022-2023 seasons.<n>Among the YOLO models, YOLOv12m achieved the best accuracy with a mAP@50 of 93.3%.
  arXiv  Detail & Related papers  (2025-09-24T21:42:24Z)
• Transfer Learning-Based CNN Models for Plant Species Identification Using Leaf Venation Patterns [0.0]
  This study evaluates the efficacy of three deep learning architectures: ResNet50, MobileNetV2, and EfficientNetB0 for automated plant species classification based on leaf venation patterns.
  arXiv  Detail & Related papers  (2025-09-03T21:23:09Z)
• Efficient Online RFT with Plug-and-Play LLM Judges: Unlocking State-of-the-Art Performance [0.0]
  Reward-model training is the cost bottleneck in modern Reinforcement Learning Human Feedback (RLHF) pipelines.<n>In the proposed method, a frozen, instruction-tuned 7B LLM is augmented with only a one line and a rank-16 LoRA adapter.<n>The plug-and-play judge 96.2% achieves accuracy on RewardBench, outperforming specialized reward networks ranging from 27B to 70B parameters.
  arXiv  Detail & Related papers  (2025-06-06T05:18:54Z)
• Common 7B Language Models Already Possess Strong Math Capabilities [61.61442513067561]
  This paper shows that the LLaMA-2 7B model with common pre-training already exhibits strong mathematical abilities. The potential for extensive scaling is constrained by the scarcity of publicly available math questions.
  arXiv  Detail & Related papers  (2024-03-07T18:00:40Z)
• Patch-Level Contrasting without Patch Correspondence for Accurate and Dense Contrastive Representation Learning [79.43940012723539]
  ADCLR is a self-supervised learning framework for learning accurate and dense vision representation. Our approach achieves new state-of-the-art performance for contrastive methods.
  arXiv  Detail & Related papers  (2023-06-23T07:38:09Z)
• Alexa Teacher Model: Pretraining and Distilling Multi-Billion-Parameter Encoders for Natural Language Understanding Systems [63.713297451300086]
  We present results from a large-scale experiment on pretraining encoders with non-embedding parameter counts ranging from 700M to 9.3B. Their subsequent distillation into smaller models ranging from 17M-170M parameters, and their application to the Natural Language Understanding (NLU) component of a virtual assistant system.
  arXiv  Detail & Related papers  (2022-06-15T20:44:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.