Upgrade to Pro — share decks privately, control downloads, hide ads and more …

CVPR2024論文紹介:Sparse Training, Continual learnin...

hinako0123
July 21, 2024
140

CVPR2024論文紹介:Sparse Training, Continual learning, Object detection

hinako0123

July 21, 2024
Tweet

Transcript

  1. Sparse trainingに関する論⽂ ・NICE: Neurogenesis Inspired Contextual Encoding for Replay- free

    Class Incremental Learning ・ MaxQ- Multi-Axis Query for N-M Sparsity Network ・Gradient-based Parameter Selection for Efficient Fine-Tuning
  2. Continual learningに関する論⽂(パラメータ⼿法) ・Gradient Reweighting: Towards Imbalanced Class-Incremental Learning ・NICE: Neurogenesis

    Inspired Contextual Encoding for Replay-free Class Incremental Learning ・Expandable Subspace Ensemble for Pre-Trained Model-Based Class-Incremental Learning ・InfLoRA: Interference-Free Low-Rank Adaptation for Continual Learning
  3. Object detectionに関する論⽂(Real-Time) ・YOLO-World- Real-Time Open-Vocabulary Object Detection ・ DETRs Beat

    YOLOs on Real-time Object Detection ・YolOOD: Utilizing Object Detection Concepts for Multi-Label Out- of-Distribution Detection ・HMD-Poser: On-Device Real-time Human Motion Tracking from Scalable Sparse Observations