Skin Cancer Captioning and VQA dataset

The original images of the datasets are from HAM10000 dataset. We have designed the captioning dataset for all the train, valid and test set from these data images and associated captions. A VQA dataset was designed from these images for test part only.  VQA dataset contains 3 types of questions. 1-what is the type of cancer?,2- what are the causes of this type of cancer? and 3-what can be the solution of this type of cancer?. The captions are in image,captions format.