Skip to content

⚡️ Speed up function get_processor by 21% #319

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

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

⚡️ Speed up function get_processor by 21% #319

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
32 changes: 18 additions & 14 deletions python/sglang/srt/utils/hf_transformers_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -422,28 +422,32 @@ def get_processor(
)

# fix: for Qwen2-VL and Sarashina2Vision models, inject default 'size' if not provided.
if config.model_type in {"qwen2_vl", "sarashina2_vision"}:
if "size" not in kwargs:
kwargs["size"] = {"shortest_edge": 3136, "longest_edge": 1003520}
if config.model_type in {"qwen2_vl", "sarashina2_vision"} and "size" not in kwargs:
# Merge directly with minimal branch logic
kwargs["size"] = {"shortest_edge": 3136, "longest_edge": 1003520}


if config.model_type not in {"llava", "clip"}:
kwargs["use_fast"] = use_fast

# Precompute all arguments for from_pretrained only once when possible
pretrained_args = {
"trust_remote_code": trust_remote_code,
"revision": revision,
**kwargs,
}
try:
if "InternVL3_5" in tokenizer_name:
processor = AutoTokenizer.from_pretrained(
tokenizer_name,
*args,
trust_remote_code=trust_remote_code,
revision=revision,
**kwargs,
**pretrained_args,
)
else:
processor = AutoProcessor.from_pretrained(
tokenizer_name,
*args,
trust_remote_code=trust_remote_code,
revision=revision,
**kwargs,
**pretrained_args,
)

except ValueError as e:
Expand All @@ -452,16 +456,16 @@ def get_processor(
logger.info(
f"Processor {tokenizer_name} does not have a slow version. Automatically use fast version"
)
kwargs["use_fast"] = True
pretrained_args["use_fast"] = True
processor = AutoProcessor.from_pretrained(
tokenizer_name,
*args,
trust_remote_code=trust_remote_code,
revision=revision,
**kwargs,
**pretrained_args,
)
else:
raise e
raise

# Efficiently process stop tokens for tokenizer
tokenizer = get_tokenizer_from_processor(processor)

attach_additional_stop_token_ids(tokenizer)
Expand Down