Hello, this topic maybe interesting for those who want to scrap images for deep learning and do it using prepared database of URLs for direct image download. For example, you may find such resource at ImageNet website in “Download” section, where you can download URLs of images in the synset.
I decided to create this piece of code: https://github.com/Vyachez/Image-scrapper-for-deep-learning to satisfy the purpose of the topic.
The code will: 1) Upload verified images (verified by size to avoid junk); 2) Create right directory structure like Jeremy advised in Lesson 1; 3) Calibrate number of images for each class at the end (equaliser.py).
This topic may look similar to this one on this forum, however focus only on image uploading from existing URLs.