We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Please download the folder code/model_training.
We have processed and labelled our images with various computer vision techniques:
4a - Context classification
4b - Human perception
4c - Instance segmentation
4d - Places365