Trending

Bookmarks

Keep Up With Latest Trending Papers. Computer Science, AI and Machine Learning and more.Subscribe

Added to collectionTop 100 Deep Learning and AI Papers 2021

Machine Learning

Computer Vision

Computation and Language

Scaling Contrastive Learning with ImageNet

Combined Scaling for Zero-shot Transfer Learning

We present a combined scaling method called basic that achieves 85.7% top-1 zero-shot accuracy on the imagenet image-text validation set, surpassing the best-published zero-shot models-clip and align-by 9.3%.

Our model also shows significant improvements in robustness benchmarks such as image-text robustness and image-text robustness.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Image-text robustness is one of the most important tasks in image processing.

Authors

Hieu Pham, Zihang Dai, Golnaz Ghiasi, Hanxiao Liu, Adams Wei Yu, Minh-Thang Luong, Mingxing Tan, Quoc V. Le

Related Topics

Contrastive learning framework

Zero-shot models

Online gradient caching

Read the Paper

Read Paper

Trending

Bookmarks

◐ Recommended Follows

Founder 42papers.com. Also bui...

Vojtěch Létal

More

◐ Latest Activity

More

Scaling Contrastive Learning with ImageNet - 42Papers

Read Related Papers

A generative adversarial network for style augmentation for nuclei segmentation

Temporal Modeling via Video Transformer and Masked Visual-token Modeling

APANet: Adaptive Prototypes Alignment Network for Few-Shot Semantic Segmentation

Data Selection Based Fine-tuning Pipeline for Image Classification

Hidden-Fold Networks: Random Recurrent Residuals Using Sparse Supermasks

Softmax Gradient Tampering: Decoupling the Backward Pass for Improved Fitting

Accelerating Deep Learning with Dynamic Data Pruning

Scaling Up Vision-Language Pre-training for Image Captioning

Self-slimmed Vision Transformer

One-shot Weakly-Supervised Segmentation in Medical Images