Skip to content

fix(tokenizer): apply config-build fallback to offset tokenizer too #75

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hallerite merged 1 commit into from
May 28, 2026
Merged

fix(tokenizer): apply config-build fallback to offset tokenizer too #75

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 6 additions & 5 deletions renderers/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1587,7 +1587,6 @@ def _get_offset_tokenizer(tokenizer):
cached = _offset_tokenizers.get(name_or_path)
if cached is not None:
return cached
from transformers import AutoTokenizer

kwargs: dict[str, Any] = {}
revision = TRUSTED_REVISIONS.get(name_or_path)
Expand All @@ -1597,10 +1596,12 @@ def _get_offset_tokenizer(tokenizer):
kwargs = {"trust_remote_code": False}
# Explicitly vanilla — we want HF's Rust tokenizer with offset
# tracking, not the fastokens shim. ``load_tokenizer`` would
# patch fastokens in by default; calling
# ``AutoTokenizer.from_pretrained`` directly here keeps the
# fastokens patch out of this code path entirely.
offset_tok = AutoTokenizer.from_pretrained(name_or_path, **kwargs)
# patch fastokens in by default; routing through
# ``_load_tokenizer_via_auto`` keeps the fastokens patch out
# of this code path while still applying the config-build
# fallback (RoPE-validation failures on nested
# ``rope_parameters``, etc.).
offset_tok = _load_tokenizer_via_auto(name_or_path, **kwargs)
if not getattr(offset_tok, "is_fast", False):
raise RuntimeError(
f"Vanilla tokenizer for {name_or_path!r} is not a fast "
Expand Down
Loading