We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
7a0e53b
Contains the weights of the model trained with multi-view supervision (this is the model presented in the paper).