PyTorch geometric train_test_split_edges implemented on dataset with edge features crashes the google-colab notebook

221 Views Asked by At

Hello stackoverflow community,

I am new to PyTorch Geometric. While I was splitting my dataset ( data size => Data(x=[14254, 1647], edge_index=[2, 8552], edge_attr=[8552, 8]) ) using train_test_split_edges from pytorch geometric the google-colab notebook crashes without any indication on the reason of this crash. My questions :

  • Why is this happening ?
  • How can I fix this ?
  • Should I implement a custom train_test_split_edges for my large dataset ?
  • How can I implement a custom train_test_split_edges ?

Best regards, Chiheb

0

There are 0 best solutions below