search

DeepGraphLearning / GearNet

GearNet and Geometric Pretraining Methods for Protein Structure Representation Learning, ICLR'2023 (https://arxiv.org/abs/2203.06125)
MIT License
253 stars 28 forks source link

TypeError: metaclass conflict: the metaclass of a derived class must be a (non-strict) subclass of the metaclasses of all its bases #56

Closed blazexpire closed 7 months ago

blazexpire commented 7 months ago

Thank you for your great work. However when try to pretrain，I encountered such error the same as a closed issue. python script/pretrain.py -c config/pretrain/mc_gearnet_edge.yaml --gpus [0] Traceback (most recent call last): File "script/pretrain.py", line 8, in from torchdrug import core, models, tasks, datasets, utils File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\torchdrug\models__init.py", line 10, in from .esm import EvolutionaryScaleModeling File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\torchdrug\models\esm.py", line 6, in import esm File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\esm__init__.py", line 8, in from .data import Alphabet, RobertaAlphabet, BatchConverter, FastaBatchedDataset # noqa File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\esm\data.py", line 11, in from torchvision.datasets.utils import download_url File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\torchvision__init__.py", line 5, in from torchvision import datasets File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\torchvision\datasets\init__.py", line 1, in from ._optical_flow import KittiFlow, Sintel, FlyingChairs, FlyingThings3D, HD1K File "C:\Users\mercy\anaconda3\envs\proj_v1\lib\site-packages\torchvision\datasets_optical_flow.py", line 26, in class FlowDataset(ABC, VisionDataset): TypeError: metaclass conflict: the metaclass of a derived class must be a (non-strict) subclass of the metaclasses of all its bases My environment is about: python 3.8.18 torch 1.12.1 torchdrug 0.2.1 torchvision 0.13.1 pyg 2.3.1 I've tried to reinstall torchdrug, but I have the same error.

Oxer11 commented 7 months ago

Hi,

This is because TorchDrug hacks the pytorch datasets' metaclass, which may lead to conflicts with datasets in torchvision. May I ask which version of esm do you use? I use fair-esm=2.0.0, which does not import torchvision and will not have this problem.

Another solution is to add from torchvision import datasets before from torchdrug import core, models, tasks, datasets, utils.

blazexpire commented 7 months ago

I've solved this problem by reinstall fair-esm. Thanks for all.