This repository was archived by the owner on Sep 9, 2026. It is now read-only.
Repository navigation
This repository was archived by the owner on Sep 9, 2026. It is now read-only.
TorchTensor cannot be deep copied. #1719
Description
Describe the bug
TorchTensor cannot be deep copied.
docarray.version
'0.36.0'
Related:
#1143
To Reproduce
script.py
import torch
from docarray import BaseDoc
from docarray.typing import TorchTensor
class Meow(BaseDoc):
embedding : TorchTensor
a = Meow(embedding=torch.randn(42))
print(id(a.embedding), id(a.copy(deep=True).embedding))Originally posted by @gusseppe in #1131 (comment)
Activity
JoanFM commentedon Jul 20, 2023
Hey @gusseppe,
What is the exact error u see?
gusseppe commentedon Jul 21, 2023
Apologize for the late response, but the error is similar to the related bug #1131:
RuntimeError: The default implementation of __deepcopy__() for non-wrapper subclasses only works for subclass types that implement new_empty() and for which that function returns another instance of the same subclass. You should either properly implement new_empty() for your subclass or override __deepcopy__() if it is intended behavior for new_empty() to return an instance of a different type.JoanFM commentedon Jul 21, 2023
We have seen it, the fix has been merged and will be released soon
gusseppe commentedon Jul 21, 2023
Super great! thanks!
JoanFM commentedon Jul 21, 2023
You can enjoy thr fix by installing the prerelease
pip install docarray --pre