File name too long error for SBMs

[GageDeZoort]

🐛 Describe the bug

If I try to create a StochasticBlockModelDataset with more than around 6-8 blocks, I get a File name too long error because the block sizes are individually copied into the filename string. I'm interested in a specific scenario where there are many blocks (think >16) that all have the same size. In this case, the processed_file_names function could just count how many blocks of the same size appear in the graph and list "degeneracy counts" next to each block size in the hash.

Environment

• PyG version: 2.1.0
• PyTorch version: 1.12.1
• OS:
• Python version: 3.10.4
• CUDA/cuDNN version:
• How you installed PyTorch and PyG (conda, pip, source): conda
• Any other relevant information (e.g., version of torch-scatter):

[rusty1s]

Thanks for reporting, I agree this isn't optimal. Do you have interest in contributing a fix?