I am having a data-set which is in tsv format. It contains website links. Where each link contain a single image. How can I read the data of more than 20,000 images without downloading each image. Help me out by giving some information that I want load the images directly from the links and to train them.
Are the links refering to the images? if yes, then images will have to be downloaded.
I am not sure what you mean by link containing images… you mean a URL contains an image? maybe you can post a sample to see what format it actually is.
Then images will have to be downloaded before they can be loaded into tensors. either 1 by 1 or download all of them at once in advance and save them locally(in the later case it will be quicker).
Imageio allows to “read” from url.