Closed ZSBSB closed 2 years ago
They are the states of the optimizers at different training stages. If the training is interrupted, you can load these optimizer states to resume training. Check code https://github.com/huzi96/Coarse2Fine-PyTorch/blob/921fcefda8205c9790c5172777a3ec371b2a6c7e/train/train.py#L685
OK, thanks for your answer.
Hello, what is the role of these three models in the training process?