Closed toledo60 closed 1 year ago
Hi, thanks for your interest in the work. The code was run with following package versions:
PyTorch - 1.9.0
GpyTorch - 1.6.0
BoTorch - 0.6.3.1
There has been new changes in later versions of GpyTorch and BoTorch which might cause some issues. However, your error looks like a dtype mismatch issue. Please try running once by converting both inputs
and train_y
to a common dtype before line 115. One way to do this is to add these two statements right after line 114:
inputs = inputs.to(torch.double)
train_y = train_y.to(torch.double)
Please let me know if it works.
Thank you, got it to work without changing anything. You don't need lines
inputs = inputs.to(torch.double)
train_y = train_y.to(torch.double)
It all just depends on the version of packages you are using, I'll later try to clean it up for newer versions.
Ive tried to run the TSP Mellows kernel example (tsp_mallow.py) and I keep receiving the following error I've tried running it locally and google Colab, different versions of gpytorch and botorch (both cpu only and gpu versions) but I keep receiving the same error every time (I haven't changed any of the code)
Im not sure what version of the packages you are using or if it even makes a difference
Could you do a pip freeze to see the package versions you used to properly run this file ? or any help is appreciated