Skip to content

Fix bos/eos tokens in bpe tokenizer in c++ #4112

@guangy10

Description

@guangy10

Those tokens are hard-coded for llama and doesn't work for other models. They should be set dynamically based on the config of the tokenizer to support HF models.

Metadata

Metadata

Assignees

Labels

module: extensionIssues related to code under extension/triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions