The easiest way to create your own world class classifiers , just type the name of anything you wish to classify between.
-
Pytorch
-
google_images_download
-
Fastai
-
Download the file and put it in your main python folder
-
now open the file and execute it
-
download any file from the internet for testing the model
-
paste the path (don't make it look like a string , no quotes needed*) and test it for yourself
Use atleast 100 images for the classification during initialization to make the model work better.more data = greater accuracy
When I tried it , the program worked really good provided just 40 images:
classified `i Phone X` and `Oneplus 6`
accuracy was 39 out of 40 times