FLAVA: A Foundational Language And Vision Alignment Model

State-of-the-art vision and vision-and-language models rely on large-scale visio-linguistic pretraining for obtaining good performance on a variety of downstream tasks. Generally, such models are often either cross-modal (contrastive) or multi-modal (with earlier fusion) but not both; and they often only target specific modalities or tasks. A promising direction would be to use a single holistic universal model, as a "foundation", that targets all modalities at once -- a true vision and language foundation model should be good at vision tasks, language tasks, and cross- and multi-modal vision and language tasks. We introduce FLAVA as such a model and demonstrate impressive performance on a wide range of 35 tasks spanning these target modalities.

PDF Abstract CVPR 2022 PDF CVPR 2022 Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Retrieval MS COCO FLAVA (zero-shot) recall@1 38.38 # 4
recall@5 67.47 # 4
Image-to-Text Retrieval MS COCO FLAVA (ViT-B, zero-shot) Recall@1 42.74 # 7
Recall@5 76.76 # 6
Zero-shot Text-to-Image Retrieval MS COCO FLAVA (ViT-B) Recall@1 38.38 # 9
Recall@5 67.47 # 7
Image Retrieval MS COCO CLIP (zero-shot) recall@1 33.29 # 5
recall@5 62.47 # 5

Methods