1Konny / FactorVAE

Pytorch implementation of FactorVAE proposed in Disentangling by Factorising(http://arxiv.org/abs/1802.05983)
MIT License
262 stars 48 forks source link

Dependency configuration and RuntimeError #13

Open jpgard opened 4 years ago

jpgard commented 4 years ago

I don't think the README lists all of the required dependencies; in particular, torchvision seems to be a required import to run the training scripts. However, torchvision is only compatible with torch 1.5 or higher, but the README lists torch 0.4 as the required version, so I doubt these are compatible.

Can you provide some more information on the exact dependency configuration required to run the scripts, perhaps in the form of a requirements.txt?

Perhaps related to this, when I attempt to reproduce the celeba results, I get the following error:

sh scripts/run_celeba.sh $RUN_NAME
...
RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [20, 256, 1, 1]] is at version 3; expected version 2 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

Is anyone else experiencing this issue?

AReinke commented 4 years ago

I receive the same runtime error.

noahcao commented 3 years ago

I'm still receiving this error. Hope someone could help to fix it.

swainsubrat commented 2 years ago

Me too!!!