Hello! Thanks for great work!!!
While reading the vita 1.5 paper, I didn't understand the part that said `Task Guidance Tokens' are trainable. Does that mean the tokens are trainable parameters? It would be helpful if you could explain this and provide some code examples for reference.
Have a good day :)
Hello! Thanks for great work!!!
While reading the vita 1.5 paper, I didn't understand the part that said `Task Guidance Tokens' are trainable. Does that mean the tokens are trainable parameters? It would be helpful if you could explain this and provide some code examples for reference.
Have a good day :)