-
Notifications
You must be signed in to change notification settings - Fork 193
Open
Labels
internal improvementsChanges that do not fix errors or add new noticeable featuresChanges that do not fix errors or add new noticeable features
Description
Recently torch 2.6 was released which changes the API for loading model weights.
This affects the the behavior of model loading for UniTS pipeline, which we should address accordingly.
UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint.
(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([scalar])` or the `torch.serialization.safe_globals([scalar])` context manager to allowlist this global if you trust this class/function.
For now, I am pinning pytorch<2.6
Metadata
Metadata
Assignees
Labels
internal improvementsChanges that do not fix errors or add new noticeable featuresChanges that do not fix errors or add new noticeable features