Skip to content

LIAFNode forward fails with missing single_step_cupy_fp32_inference attribute #655

@GraceDyer

Description

@GraceDyer

Issue type

  • Bug Report
  • Feature Request
  • Help wanted
  • Other

SpikingJelly version

0.0.0.0.14

Description

I encountered a reproducible issue where LIAFNode fails during forward execution with a missing attribute error.

This appears to affect normal forward usage rather than a corner-case export or serialization workflow.

Observed error:

AttributeError: 'LIAFNode' object has no attribute 'single_step_cupy_fp32_inference'

This behavior was reproducible across repeated reruns in my validation.

Minimal code to reproduce the error/bug

import torch
from spikingjelly.activation_based import neuron

x = torch.randn(2, 4)
node = neuron.LIAFNode()

node(x)

Expected behavior

LIAFNode should complete the forward pass successfully, or fail with a clearly documented limitation if additional setup is required.

Environment

  • torch: 2.8.0
  • spikingjelly: 0.0.0.0.14
  • python: 3.10

Additional notes

This issue matters because it breaks the normal forward path for LIAFNode.

If this attribute is expected to be initialized only under certain conditions, an explicit validation check or clearer documentation would still be very helpful.

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions