Search Results for author:
Found 7 papers, 6 papers with code
vid-TLDR: Training Free Token merging for Light-weight Video Transformer
To tackle these issues, we propose training free token merging for lightweight video Transformer (vid-TLDR) that aims to enhance the efficiency of video Transformers by merging the background tokens without additional training.
Ranked #2 on
Video Retrieval
on SSv2-template retrieval
(using extra training data)
Multi-criteria Token Fusion with One-step-ahead Attention for Efficient Vision Transformers
Here, we argue that token fusion needs to consider diverse relations between tokens to minimize information loss.
Ranked #1 on
Efficient ViTs
on ImageNet-1K (With LV-ViT-S)
Read-only Prompt Optimization for Vision-Language Few-shot Learning
RPO leverages masked attention to prevent the internal representation shift in the pre-trained model.
Ranked #6 on
Prompt Engineering
on Caltech-101
Concept Bottleneck with Visual Concept Filtering for Explainable Medical Image Classification
However, those methods do not consider whether a concept is visually relevant or not, which is an important factor in computing meaningful concept scores.
MELTR: Meta Loss Transformer for Learning to Fine-tune Video Foundation Models
Therefore, we propose MEta Loss TRansformer (MELTR), a plug-in module that automatically and non-linearly combines various loss functions to aid learning the target task via auxiliary learning.
Ranked #2 on
Video Captioning
on YouCook2
TokenMixup: Efficient Attention-guided Token-level Data Augmentation for Transformers
To this end, we propose TokenMixup, an efficient attention-guided token-level data augmentation method that aims to maximize the saliency of a mixed set of tokens.
Ranked #65 on
Image Classification
on CIFAR-10
Video-Text Representation Learning via Differentiable Weak Temporal Alignment
In this paper, we propose a novel multi-modal self-supervised framework Video-Text Temporally Weak Alignment-based Contrastive Learning (VT-TWINS) to capture significant information from noisy and weakly correlated data using a variant of Dynamic Time Warping (DTW).
