Skip to content
This repository has been archived by the owner on May 19, 2023. It is now read-only.

[FEA] Support different models in sequence classifier #381

Open
efajardo-nv opened this issue Apr 8, 2021 · 2 comments
Open

[FEA] Support different models in sequence classifier #381

efajardo-nv opened this issue Apr 8, 2021 · 2 comments

Comments

@efajardo-nv
Copy link
Contributor

Is your feature request related to a problem? Please describe.
Sequence classifier only uses bert-base-uncased base model and corresponding vocabulary and tokenizer for fine tuning.

Describe the solution you'd like
Allow use of different base models and way to pass corresponding vocabulary hash file for cudf subword tokenizer.

@efajardo-nv efajardo-nv added the feature request New feature or request label Apr 8, 2021
@github-actions
Copy link

github-actions bot commented May 8, 2021

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.

@github-actions
Copy link

This issue has been labeled inactive-90d due to no recent activity in the past 90 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
Status: No status
Development

No branches or pull requests

1 participant