Skip to content

Loss and labels #15

Description

@DebanjanaKar

In the paper, in section 3, there is a mention of the loss function only attending to the utterances and not the intent during training. "During training, the prediction loss is only computed for x0, and A0 provides the attended conditions." But I look at the labels provided for training, it includes the intent as well. Can you please help me understand how the code achieves to compute loss only for the utterance leaving out the intent? Any help would really be appreciated.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions