BenthicNet Pre-trained Models
Collection
BenthicNet pre-trained (self-supervised) computer vision models
•
5 items
•
Updated
Pre-training Dataset: BenthicNet (Training)
Method: MoCo-v3
Architecture: ViT-B
Length: 100 Epochs
BenthicNet: https://arxiv.org/abs/2405.05241
Code to construct the dataset: https://github.com/DalhousieAI/pangaea-downloader