![Mitchell Wortsman on X: "Can zero-shot models such as CLIP be fine-tuned without reducing out-of-distribution accuracy? Yes! Our new method for robust fine-tuning improves average OOD accuracy by 9% on multiple ImageNet Mitchell Wortsman on X: "Can zero-shot models such as CLIP be fine-tuned without reducing out-of-distribution accuracy? Yes! Our new method for robust fine-tuning improves average OOD accuracy by 9% on multiple ImageNet](https://pbs.twimg.com/media/E-yNT4JUcAUFAoa.jpg:large)
Mitchell Wortsman on X: "Can zero-shot models such as CLIP be fine-tuned without reducing out-of-distribution accuracy? Yes! Our new method for robust fine-tuning improves average OOD accuracy by 9% on multiple ImageNet
ZeroForge Block Diagram. Illustration of our novel method for learning... | Download Scientific Diagram
GitHub - openai/CLIP: CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
![Zero-shot top1 accuracy on ImageNet. All the models are pre-trained for... | Download Scientific Diagram Zero-shot top1 accuracy on ImageNet. All the models are pre-trained for... | Download Scientific Diagram](https://www.researchgate.net/publication/360310696/figure/tbl1/AS:1151124153344006@1651460793472/Zero-shot-top1-accuracy-on-ImageNet-All-the-models-are-pre-trained-for-32-epochs.png)
Zero-shot top1 accuracy on ImageNet. All the models are pre-trained for... | Download Scientific Diagram
![Expert-level detection of pathologies from unannotated chest X-ray images via self-supervised learning | Nature Biomedical Engineering Expert-level detection of pathologies from unannotated chest X-ray images via self-supervised learning | Nature Biomedical Engineering](https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41551-022-00936-9/MediaObjects/41551_2022_936_Fig1_HTML.png)
Expert-level detection of pathologies from unannotated chest X-ray images via self-supervised learning | Nature Biomedical Engineering
The overall framework of the proposed TAP-C method for zero-shot VQA.... | Download Scientific Diagram
![CLIP-Sculptor: Zero-Shot Generation of High-Fidelity and Diverse Shapes from Natural Language | Brown IVL CLIP-Sculptor: Zero-Shot Generation of High-Fidelity and Diverse Shapes from Natural Language | Brown IVL](https://ivl.cs.brown.edu/assets/images/projects/clip-sculptor/method_diagram.jpg)
CLIP-Sculptor: Zero-Shot Generation of High-Fidelity and Diverse Shapes from Natural Language | Brown IVL
![Using CLIP to Classify Images without any Labels | by Cameron R. Wolfe, Ph.D. | Towards Data Science Using CLIP to Classify Images without any Labels | by Cameron R. Wolfe, Ph.D. | Towards Data Science](https://miro.medium.com/v2/resize:fit:1400/1*JnDxBVCAu3S4vLhM0jVw4g.png)
Using CLIP to Classify Images without any Labels | by Cameron R. Wolfe, Ph.D. | Towards Data Science
![Using CLIP to Classify Images without any Labels | by Cameron R. Wolfe, Ph.D. | Towards Data Science Using CLIP to Classify Images without any Labels | by Cameron R. Wolfe, Ph.D. | Towards Data Science](https://miro.medium.com/v2/resize:fit:1400/1*XjDmui7Zprb809NsMNBV4Q.png)