PaperICML 2021 · 2021
Learning Transferable Visual Models From Natural Language Supervision
1page built on it
0more pages cite it
1narrative
3registry papers cite it
In the Atlas
ModelPrimary source of
CLIP
Contrastive image–text pretraining on 400M web pairs: dual encoders in one embedding space make natural language the classifier — zero-shot ImageNet at 76.2% via prompts, with unmatched robustness under distribution shift.
Appears in 1 narrative
Lineage
Papers in the Atlas registry only · one dot per paperBuilt upon by · 3
Builds on · 4
2016Deep Residual Learning for Image RecognitionHe, Zhang et al.2017Attention Is All You NeedVaswani, Shazeer et al.2020A Simple Framework for Contrastive Learning of Visual RepresentationsChen, Kornblith et al.2020An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleDosovitskiy, Beyer et al.Built upon by · 3
2023WinCLIP: Zero-/Few-Shot Anomaly Classification and SegmentationJeong, Zou et al.2023AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly DetectionZhou, Pang et al.2025SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense FeaturesTschannen, Gritsenko et al.