Revisiting Classifier: Transferring Vision-Language Models for Video Recognition

4 Jul 2022  ·  Wenhao Wu, Zhun Sun, Wanli Ouyang ·

Transferring knowledge from task-agnostic pre-trained deep models for downstream tasks is an important topic in computer vision research. Along with the growth of computational capacity, we now have open-source vision-language pre-trained models in large scales of the model architecture and amount of data. In this study, we focus on transferring knowledge for video classification tasks. Conventional methods randomly initialize the linear classifier head for vision classification, but they leave the usage of the text encoder for downstream visual recognition tasks undiscovered. In this paper, we revise the role of the linear classifier and replace the classifier with the different knowledge from pre-trained model. We utilize the well-pretrained language model to generate good semantic target for efficient transferring learning. The empirical study shows that our method improves both the performance and the training speed of video classification, with a negligible change in the model. Our simple yet effective tuning paradigm achieves state-of-the-art performance and efficient training on various video recognition scenarios, i.e., zero-shot, few-shot, general recognition. In particular, our paradigm achieves the state-of-the-art accuracy of 87.8% on Kinetics-400, and also surpasses previous methods by 20~50% absolute top-1 accuracy under zero-shot, few-shot settings on five popular video datasets. Code and models can be found at https://github.com/whwu95/Text4Vis .

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Action Recognition ActivityNet Text4Vis (w/ ViT-L) mAP 96.9 # 1
Zero-Shot Action Recognition ActivityNet Text4Vis Top-1 Accuracy 84.6 # 2
Zero-Shot Action Recognition HMDB51 Text4Vis Top-1 Accuracy 58.4 # 7
Zero-Shot Action Recognition Kinetics Text4Vis Top-1 Accuracy 68.9 # 6
Top-5 Accuracy 90.3 # 3
Action Classification Kinetics-400 Text4Vis (CLIP ViT-L/14) Acc@1 87.8 # 25
Acc@5 97.6 # 18
Zero-Shot Action Recognition UCF101 Text4Vis Top-1 Accuracy 85.8 # 6
Action Recognition UCF101 Text4Vis 3-fold Accuracy 98.2 # 8

Methods