
MedCLIP is a Python package for contrastive learning from unpaired medical images and texts, published at EMNLP 2022. It provides pretrained vision-text models loadable in a few lines.
Web appFree
Step 1: Pick a PyTorch version with GPU support from pytorch.org.
Step 2: Run pip install medclip, or install from the GitHub repository.
Step 3: Load a model with MedCLIPModel and from_pretrained.
Step 4: Prepare an image and text with MedCLIPProcessor.
Step 5: Run the model, or wrap it in PromptClassifier for classification.
MedCLIP-ResNet50 and MedCLIP-ViT, selected through the vision_cls argument.
The PromptClassifier compares an image with class prompts and returns logits per class, shown with CheXpert-style classes.
The README refers users to the Stanford CheXpert labeler for information extraction tools.
Cite Wang, Wu, Agarwal and Sun, MedCLIP: Contrastive Learning from Unpaired Medical Images and Texts, EMNLP 2022.
FloraceneAI app builder for creating and deploying HIPAA-compliant internal healthcare tools without a dedicated engineering team.
BiomedCLIPBiomedCLIP is a biomedical vision-language model from Microsoft, pretrained on 15 million figure-caption pairs from PubMed Central. It supports retrieval, image classification and visual question answering research.MONAI Generative ModelsOpen medical imaging AI framework for generative models, synthetic imaging, segmentation, and healthcare research