01_pytorch_workflow.ipynb plot_predictions fails with pytorch 2.3.1

[starksm64]

I have cloned the https://github.com/mrdbourke/pytorch-deep-learning.git repo and was looking at the 01_pytorch_workflow.ipynb after having gone through the 00_pytorch_fundamentals.ipynb notebook.

Early on when the plot_predictions function is first used, I see this error:
ValueError: object __array__ method not producing an array

torch.version in my environment is '2.3.1.post100'

Looking at the matlibplot.scatter function defintion, it says x should be a float or ArrayLike. Apparently you cannot pass in a tensor as the notebook is doing. Switching the first example usage of the plot_predictions function to convert the tensors to numpy arrays like this fixes the error:

plot_predictions(train_data=X_train.numpy(), train_labels=Y_train.numpy(), test_data=X_test.numpy(), test_labels=Y_test.numpy())

Does the course require a 1.x version of pytorch?

[ReactjsGod]

Thanks for sharing this @starksm64 very helpful.