Open QuinnHe opened 2 weeks ago
I wonder if something like this would be more accurate?
The ml5.js image classifier uses the pre-trained MobileNet model by default. You can optionally load and use other models such as Darknet as well as a custom-trained model, DoodleNet, which is also built upon the MobileNet architecture and trained on images from the Google Quick, Draw dataset.
To-do As Dan mentioned above, go through all model sources, and provide a bit more detailed info.
For instance, "The ml5.js image classifier is based on the MobileNet, Darknet, and DoodleNet." We need to go over each model, and check whether the source link is correct.