We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
810d29a
The archive contains the weights for models trained 4 epochs on ImageNet with image size 256x256, batch size 8 and L1 or L2 loss.
