Skip to content
This repository was archived by the owner on Jun 3, 2025. It is now read-only.

Export files required by OPT to load the tokenizer #1571

Merged
merged 5 commits into from
May 22, 2023
Merged

Export files required by OPT to load the tokenizer #1571

merged 5 commits into from
May 22, 2023

Conversation

dbogunowicz
Copy link
Contributor

OPT does not contain a "tokenizer.json", but instead requires the following files:
"special_tokens_map.json"
"vocab.json"
"merges.txt"
to create the tokenizer from pretrained.

I.e if those three files are present in the deployment model this line of code will execute:

model_path = "deployment"
tokenizer = AutoTokenizer.from_pretrained(
            model_path,
        )

@dbogunowicz dbogunowicz requested review from bfineran, natuan, a team and rahul-tuli and removed request for a team May 19, 2023 15:59
natuan
natuan suggested changes May 19, 2023
View reviewed changes
@dbogunowicz dbogunowicz requested a review from natuan May 22, 2023 08:31
anmarques
anmarques reviewed May 22, 2023
View reviewed changes
dbogunowicz
dbogunowicz commented May 22, 2023
View reviewed changes
@dbogunowicz dbogunowicz requested a review from anmarques May 22, 2023 15:06
@natuan natuan merged commit 3dd1f8d into main May 22, 2023
@natuan natuan deleted the feature/damian/tokenizer_files branch May 22, 2023 15:32
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@bfineran bfineran Awaiting requested review from bfineran

@rahul-tuli rahul-tuli Awaiting requested review from rahul-tuli rahul-tuli was automatically assigned from neuralmagic/machine-learning

2 more reviewers

@natuan natuan natuan approved these changes

@anmarques anmarques anmarques approved these changes

Reviewers whose approvals may not affect merge requirements
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@dbogunowicz @natuan @bfineran @anmarques