Skip to content

Avoid calls to tokenizer.added_tokens_decoder #12473

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
merged 1 commit into from
Mar 20, 2025
Merged

Avoid calls to tokenizer.added_tokens_decoder #12473

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
6 changes: 4 additions & 2 deletions convert_hf_to_gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -529,6 +529,8 @@ def get_vocab_base(self) -> tuple[list[str], list[int], str]:
reverse_vocab = {id_: encoded_tok for encoded_tok, id_ in tokenizer.vocab.items()}
added_vocab = tokenizer.get_added_vocab()

added_tokens_decoder = tokenizer.added_tokens_decoder

for i in range(vocab_size):
if i not in reverse_vocab:
tokens.append(f"[PAD{i}]")
Expand All @@ -538,13 +540,13 @@ def get_vocab_base(self) -> tuple[list[str], list[int], str]:
if token in added_vocab:
# The tokenizer in llama.cpp assumes the CONTROL and USER_DEFINED tokens are pre-normalized.
# To avoid unexpected issues - we make sure to normalize non-normalized tokens
if not tokenizer.added_tokens_decoder[i].normalized:
if not added_tokens_decoder[i].normalized:
previous_token = token
token = tokenizer.decode(tokenizer.encode(token, add_special_tokens=False))
if previous_token != token:
logger.info(f"{repr(previous_token)} is encoded and decoded back to {repr(token)} using AutoTokenizer")

if tokenizer.added_tokens_decoder[i].special or self.does_token_look_special(token):
if added_tokens_decoder[i].special or self.does_token_look_special(token):
toktypes.append(gguf.TokenType.CONTROL)
else:
# NOTE: this was added for Gemma.
Expand Down