Calling AutoModel.from_config() method for a model requiring timm does not raise ImportError although it should #20671
Description
System Info
transformersversion: 4.26.0.dev0- Platform: Linux-5.15.0-56-generic-x86_64-with-glibc2.35
- Python version: 3.9.12
- Huggingface_hub version: 0.11.0.dev0
- PyTorch version (GPU?): 1.12.1+cu102 (True)
- Tensorflow version (GPU?): 2.9.1 (True)
- Flax version (CPU?/GPU?/TPU?): 0.5.2 (cpu)
- Jax version: 0.3.14
- JaxLib version: 0.3.14
- Using GPU in script?:
- Using distributed or parallel set-up in script?:
Who can help?
No response
Information
- The official example scripts
- My own modified scripts
Tasks
- An officially supported task in the
examplesfolder (such as GLUE/SQuAD, ...) - My own task or dataset (give details below)
Reproduction
pip uninstall timm, and then:
from transformers import AutoModel, AutoConfig
cfg = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-detr")
model = AutoModel.from_config(cfg)raising:
Traceback (most recent call last):
File "<tmp 1>", line 18, in <module>
model = AutoModel.from_config(cfg)
File "/home/fxmarty/hf_internship/transformers/src/transformers/models/auto/auto_factory.py", line 410, in from_config
return model_class._from_config(config, **kwargs)
File "/home/fxmarty/hf_internship/transformers/src/transformers/utils/import_utils.py", line 1008, in __getattribute__
return super().__getattribute__(key)
AttributeError: type object 'DetrModel' has no attribute '_from_config'
Expected behavior
It should raise:
ImportError:
DetrModel requires the timm library but it was not found in your environment. You can install it with pip:
`pip install timm`. Please note that you may need to restart your runtime after installation.