How can I use your open-source model to perform without training the encoder and decoder?
How can I use your open-source model to perform without training the encoder and decoder?