Generate serialized data in build script #24
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Instead of including the serialized data in the crate, @aneubeck suggested that we generate it at build time.
This PR introduces a new
bpe-openai
crate that generates serialized data as part of its build. This cannot be part ofbpe
directly, because that would requirebpe
to be available whenbpe
is built. Thebpe-openai
crate exposed public functions to get static BPE instances loaded from the serialized data generated at build time. As a convenience, it re-exports thebpe
crate, so depending on this crate is enough to be able to use these tokenizers fully.Serializing the data is a hit on compile time. But I think this shouldn't happen too often with the
rerun-if-change:build.rs
directive.