Skip to content

The exception of consistency loss #14

@lixiangcog

Description

@lixiangcog

During the training process, the consistency loss is getting bigger and bigger, and the other losses are getting smaller and smaller. Is it normal? Is it what we expect?

            consistency_loss = 0
            consistency_weight = get_current_consistency_weight(iter_num//len(loader_train_s), max_epoch)
            consistency_dist = consistency_criterion(predout_t[train_params['labeled_bs']:], ema_output)    #(batch, 3, 256, 256)
            consistency_dist = torch.mean(consistency_dist)
            consistency_loss = consistency_dist * consistency_weight

🐱‍👤

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions