Skip to content

When training ranking models, should we use explicit negative samples or negative samples sampled from the batch, or provided by the retrieval model? #669

Open
Open
When training ranking models, should we use explicit negative samples or negative samples sampled from the batch, or provided by the retrieval model?#669
Assignees
gabrielspmoreirasararb
Milestone
Merlin 22.12

Description

@viswa-nvidia
viswa-nvidia
opened on Oct 12, 2022
No description provided.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions