Skip to content

sparse tensors not supported by default collate_fn #1

@daskandalis

Description

@daskandalis

When training with 02_train_network.py, execution terminates with the error:

RuntimeError: Batches of sparse tensors are not currently supported by the default collate_fn; please provide a custom collate_fn to handle them appropriately.

I notice that train_loader is not passed collate_fn when batch_size==1.

This is on pytorch 2.4.0 and CUDA 12.4, which otherwise works up to that point.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions