Skip to content

Assemble big docket entry dataset #96

@nadahlberg

Description

@nadahlberg

In early experiments the 6.5M entries is not enough data for models to converge. Even with up to 5 epochs. Pull a much larger set of data (from the dev db) for the "real" training runs.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

Status

In progress

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions