issue with new transformer
#38
by
ABDALLALSWAITI
- opened
bash
model = AutoModelForImageSegmentation.from_pretrained('briaai/RMBG-2.0', trust_remote_code=True)
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/models/auto/auto_factory.py", line 597, in from_pretrained
return model_class.from_pretrained(
~~~~~~~~~~~~~~~~~~~~~~~~~~~^
pretrained_model_name_or_path, *model_args, config=config, **hub_kwargs, **kwargs
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/modeling_utils.py", line 288, in _wrapper
return func(*args, **kwargs)
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/modeling_utils.py", line 5106, in from_pretrained
model.tie_weights()
~~~~~~~~~~~~~~~~~^^
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/modeling_utils.py", line 3136, in tie_weights
module.tie_embeddings_and_encoder_decoder()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/modeling_utils.py", line 3112, in tie_embeddings_and_encoder_decoder
if getattr(self.config.get_text_config(decoder=True), "tie_word_embeddings", True):
~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/configuration_utils.py", line 1238, in get_text_config
if not return_both and len(valid_text_config_names) == 0 and config_to_return.is_encoder_decoder:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/abdallah/Desktop/test imagen/.venv/lib/python3.13/site-packages/transformers/configuration_utils.py", line 207, in __getattribute__
return super().__getattribute__(key)
~~~~~~~~~~~~~~~~~~~~~~~~^^^^^
AttributeErro
worked with transformers==4.54.1
I copied an updated birefnet.py (https://huggingface.co/ZhengPeng7/BiRefNet/blob/main/birefnet.py) into the RMBG model folder and it appears to work.
