Skip to content

Use slow tokenizer for open llama models #168

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
Jun 20, 2023
Merged

Use slow tokenizer for open llama models #168

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
7 changes: 6 additions & 1 deletion vllm/engine/tokenizer_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,12 @@ def get_tokenizer(
) -> Union[PreTrainedTokenizer, PreTrainedTokenizerFast]:
"""Gets a tokenizer for the given model name via Huggingface."""
config = AutoConfig.from_pretrained(model_name)
if config.model_type == "llama" and getattr(kwargs, "use_fast", True):
if "open_llama" in model_name:
kwargs["use_fast"] = False
logger.info(
"OpenLLaMA models do not support the fast tokenizer. "
"Using the slow tokenizer instead.")
elif config.model_type == "llama" and getattr(kwargs, "use_fast", True):
# LLaMA fast tokenizer causes protobuf errors in some environments.
# However, we found that the below LLaMA fast tokenizer works well in
# most environments.
Expand Down