Files
vllm/vllm/transformers_utils/image_processor.py
Cyrus Leung 9831aec49f [Core] Dynamic image size support for VLMs (#5276)
Signed-off-by: Xiaowei Jiang <xwjiang2010@gmail.com>
Co-authored-by: Xiaowei Jiang <xwjiang2010@gmail.com>
Co-authored-by: ywang96 <ywang@roblox.com>
Co-authored-by: xwjiang2010 <87673679+xwjiang2010@users.noreply.github.com>
Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>
2024-07-02 20:34:00 -07:00

38 lines
1.4 KiB
Python

from typing import cast
def get_image_processor(
processor_name: str,
*args,
trust_remote_code: bool = False,
**kwargs,
):
"""Gets an image processor for the given model name via HuggingFace."""
# don't put this import at the top level
# it will call torch.cuda.device_count()
from transformers import AutoImageProcessor
from transformers.image_processing_utils import BaseImageProcessor
try:
processor = AutoImageProcessor.from_pretrained(
processor_name,
*args,
trust_remote_code=trust_remote_code,
**kwargs)
except ValueError as e:
# If the error pertains to the processor class not existing or not
# currently being imported, suggest using the --trust-remote-code flag.
# Unlike AutoTokenizer, AutoImageProcessor does not separate such errors
if not trust_remote_code:
err_msg = (
"Failed to load the image processor. If the image processor is "
"a custom processor not yet available in the HuggingFace "
"transformers library, consider setting "
"`trust_remote_code=True` in LLM or using the "
"`--trust-remote-code` flag in the CLI.")
raise RuntimeError(err_msg) from e
else:
raise e
return cast(BaseImageProcessor, processor)