AdaptiveAvgPool1d与BatchNorm1在多步传播模式下报错

[Zihao0]

Read before creating a new issue

• Users who want to use SpikingJelly should first be familiar with the usage of PyTorch.
• If you do not know much about PyTorch, we recommend that the user can learn the basic tutorials of PyTorch.
• Do not ask for help with the basic conception of PyTorch/Machine Learning but not related to SpikingJelly. For these questions, please refer to Google or PyTorch Forums.

For faster response

You can @ the corresponding developers for your issue. Here is the division:

Features	Developers
Neurons and Surrogate Functions	fangwei123456 Yanqi-Chen
CUDA Acceleration	fangwei123456 Yanqi-Chen
Reinforcement Learning	lucifer2859
ANN to SNN Conversion	DingJianhao Lyu6PosHao
Biological Learning (e.g., STDP)	AllenYolk
Others	Grasshlw lucifer2859 AllenYolk Lyu6PosHao DingJianhao Yanqi-Chen fangwei123456

We are glad to add new developers who are volunteering to help solve issues to the above table.

Issue type

• Bug Report
• Feature Request
• Help wanted
• Other

SpikingJelly version

0.0.0.0.14

Description

...

Minimal code to reproduce the error/bug

import spikingjelly
# ...

import numpy as np
import torch
from spikingjelly.activation_based import neuron, layer, functional
import torch
import torch.nn as nn


class CustomModel(nn.Module):
    def __init__(self, num_cls):
        super(CustomModel, self).__init__()
        self.T = 3
        # Layer 3
        self.fc1 = nn.Sequential(
            # nn.AdaptiveAvgPool1d(num_cls),
            layer.BatchNorm1d(20)
            # nn.Flatten(),  # 添加展平层
            # nn.Linear(64 * 6 * 6, num_cls * 10),
        )

        functional.set_step_mode(self, step_mode='m')

    def forward(self, x):
        batch_size = x.size()[0]
        # encoder layer
        x = x.unsqueeze(0).repeat(self.T, 1, 1)
        print(f"Input shape: {x.shape}")
        x = self.fc1(x)
        print(f"After fc1 shape: {x.shape}")
        return x

model = CustomModel(num_cls=2)
input_tensor = torch.randn(256, 20)  # 示例输入张量
output = model(input_tensor)
print(output.shape)  # 输出

报错

ValueError: expected x with shape [T, N, C, L], but got x with shape torch.Size([3, 256, 20])!

问题

奇怪的是，在单步模式下，输入大小为[N, C]，程序可以正常运行，多步模式下将输入扩展为[T, N, C]就会发生上述报错，提示期望的输入大小为 [T, N, C, L]，这是为什么呢，不应该是 [T, N, C]吗？单步模式下的输入为[N, C]，多步不应该为[T, N, C]吗？
上述问题在AdaptiveAvgPool1d与BatchNorm1d都存在，单步可以运行，但多步需要多一个L维度，请问这是为什么？怎样解决？谢谢！

[Zihao0]

问题已经解决了，如果作者愿意解释一下就更好了