Skip to content

two training loss per step #46

@fzuo1230

Description

@fzuo1230

hello, i was training wan2.1-14b-i2v-720p, but i found that i get two loss print after one step:

Average Epoch time: 231.10 seconds
Average Peak memory 68550.57MiB

Epoch 0: 1%|▏ | 3/217 [13:05<15:33:34, 0.00it/s, v_num=0, train_loss=0.187]
Epoch 0: 1%|▏ | 3/217 [13:05<15:33:34, 0.00it/s, v_num=0, train_loss=0.118]

Is this normal or am I doing something wrong?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions