TODO list

The Road Less Scheduled

ImageBind: One Embedding Space To Bind Them All Group ViT

beyond sight: https://arxiv.org/pdf/2501.04693

SSL 표준

• [moco1234, simclr]
• SwAV: Unsupervised Learning of Visual Features by Contrasting Cluster Assignments
• Simsiam: Exploring Simple Siamese Representation Learning
• CPCv2: Data-Efficient Image Recognition with Contrastive Predictive Coding
• CPC: Representation Learning with Contrastive Predictive Coding

Transformer Scaling방법들

• Scaling vision transformers
• OpenCLIP: Robust fine-tuning of zero-shot models

Transformer의 Token 변형

• Memory Transformer
• Recurrent Memory Transformer
• Fine-tuning Image Transformers using Learnable Memory

Transformer Positional Encoding

• [NTK: ]
• [푸리에]

Transformer의 Complexity

• [linformer]

Detection

• Localizing Objects with Self-Supervised Transformers and no Labels
• You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection
• ViDT: An Efficient and Effective Fully Transformer-based Object Detector