Improving multimodal datasets with image captioning

Open in new window