The settings of training parameters #2
Description
Hello, dear author. I would like to inquire about the training parameters mentioned in your paper, specifically regarding the learning rate and batch size settings. The paper states that the optimizer used for training is Adam, trained over 100 epochs, with an initial learning rate of 0.0001, which is adjusted to 0.003 after 50 epochs. Does this mean that a learning rate of 0.0001 is used for the first 50 epochs, and 0.003 for the remaining 50 epochs?
Additionally, regarding the batch size, the paper mentions a batch size of 2, but the table includes both 2 and 4 as batch size values. Could you clarify the exact batch size setting? If hardware permits, would you recommend using a larger batch size, such as 24 or 32? I would also like to know if this would impact model training.
Best regards,