Building models that can be rapidly adapted to numerous tasks using only a handful of annotated examples is an open challenge for multimodal machine learning research. We introduce Flamingo, a family of Visual Language Models (VLM) with this ability. Flamingo models include key architectural innovations to: (i) bridge powerful pretrained vision-only and language-only models, (ii) handle sequences of arbitrarily interleaved visual and textual data, and (iii) seamlessly ingest images or videos as inputs. Thanks to their flexibility, Flamingo models can be trained on large-scale multimodal web corpora containing arbitrarily interleaved text and images, which is key to endow them with in-context few-shot learning capabilities. We perform a thorough evaluation of the proposed Flamingo models, exploring and measuring their ability to rapidly adapt to a variety of image and video understanding benchmarks. These include open-ended tasks such as visual question-answering, where the model is prompted with a question which it has to answer, captioning tasks, which evaluate the ability to describe a scene or an event, and close-ended tasks such as multiple choice visual question-answering. For tasks lying anywhere on this spectrum, we demonstrate that a single Flamingo model can achieve a new state of the art for few-shot learning, simply by prompting the model with task-specific examples. On many of these benchmarks, Flamingo actually surpasses the performance of models that are fine-tuned on thousands of times more task-specific data.

PDF Abstract DeepMind 2022 PDF

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Zero-Shot Cross-Modal Retrieval COCO 2014 Flamingo Image-to-text R@1 65.9 # 4
Image-to-text R@5 87.3 # 3
Image-to-text R@10 92.9 # 3
Text-to-image R@1 48.0 # 4
Text-to-image R@5 73.3 # 4
Text-to-image R@10 82.1 # 3
Zero-Shot Cross-Modal Retrieval Flickr30k Flamingo Image-to-text R@1 89.3 # 5
Image-to-text R@5 98.8 # 4
Image-to-text R@10 99.7 # 2
Text-to-image R@1 79.5 # 3
Text-to-image R@5 95.3 # 2
Text-to-image R@10 97.9 # 1
Meme Classification Hateful Memes Flamingo (few-shot:32) ROC-AUC 0.700 # 4
Zero-Shot Learning iVQA Flamingo (0-shot) Accuracy 0.407 # 1
Visual Question Answering MSRVTT-QA Flamingo (32-shot) Accuracy 0.310 # 10
Visual Question Answering MSRVTT-QA Flamingo (0-shot) Accuracy 0.174 # 12
Temporal/Casual QA NExT-QA Flamingo WUPS 33.5 # 1
Action Recognition RareAct 🦩 Flamingo mWAP 60.8 # 1

Methods


No methods listed for this paper. Add relevant methods here