parameter-efficient adaptation
This approach strives to fine-tune models with minimal adjustments to their parameters. It allows practitioners to leverage large pre-trained models' capabilities without extensive retraining, often improving efficiency in scenarios with constrained resources.
- EditInfinity: Image Editing with Binary-Quantized Generative Models
- LoRA-EnVar: Parameter-Efficient Hybrid Ensemble Variational Assimilation for Weather Forecasting
- SEMPO: Lightweight Foundation Models for Time Series Forecasting
- Test-Time Adaptive Object Detection with Foundation Model
- VESSA: Video-based objEct-centric Self-Supervised Adaptation for Visual Foundation Models