docs: fix typo in convert.py (#17)

signficant -> significant
This commit is contained in:
Ikko Eltociear Ashimine 2023-08-21 22:56:21 +09:00 committed by GitHub
parent a22722631a
commit cb01adee20
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -111,7 +111,7 @@ def bytes_to_unicode():
The reversible bpe codes work on unicode strings.
This means you need a large # of unicode characters in your vocab if you want to avoid UNKs.
When you're at something like a 10B token dataset you end up needing around 5K for decent coverage.
This is a signficant percentage of your normal, say, 32K bpe vocab.
This is a significant percentage of your normal, say, 32K bpe vocab.
To avoid that, we want lookup tables between utf-8 bytes and unicode strings.
And avoids mapping to whitespace/control characters the bpe code barfs on.
"""