Open jingru-lin opened 1 year ago
According to the paper, the contrastive loss weight is set to 1e-5 * num_train_steps.
However, in the given config, the 'loss_weights: [10,1]', where 1 is the contrastive loss weight.
May I know why?
The run.sh will override the config.
According to the paper, the contrastive loss weight is set to 1e-5 * num_train_steps.
However, in the given config, the 'loss_weights: [10,1]', where 1 is the contrastive loss weight.
May I know why?