I tried to modify ResNet18 using CKAN, but encountered a gradient computation failure issue

[icadada]

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [2, 512, 7, 7]], which is output 0 of ReluBackward0, is at version 3; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

This error message indicates that a variable was modified in-place during gradient computation, leading to the gradient computation failure. The in-place operation in the ReLU of ResNet18 has already been set to False, so it is suspected that the in-place operation is caused by CKAN.

"My complete code is as follows.

import torch.nn as nn
from kan_convolutional.KANConv import KAN_Convolutional_Layer
from kan import KAN
import math



class BasicBlock(nn.Module):
    def __init__(self, in_channels, out_channels, stride=1, downsample=None):
        super(BasicBlock, self).__init__()
        self.in_channel = in_channels
        self.out_channel = out_channels
        # self.conv1 = nn.Conv2d(in_channels, out_channels, 3, stride, padding=1, bias=False)
        self.conv1 = KAN_Convolutional_Layer(
            n_convs=int(out_channels//in_channels),
            kernel_size=(3, 3),
            stride=(stride, stride),
            padding=(1, 1),
            device='cuda'
        )
        self.bn1 = nn.BatchNorm2d(out_channels)
        # self.conv2 = nn.Conv2d(out_channels, out_channels, 3, 1, padding=1, bias=False)
        self.conv2 = KAN_Convolutional_Layer(
            n_convs=int(out_channels//out_channels),
            kernel_size=(3, 3),
            stride=(1, 1),
            padding=(1, 1),
            device='cuda'
        )
        self.bn2 = nn.BatchNorm2d(out_channels)
        self.relu = nn.ReLU(inplace=True)
        self.downsample = downsample

    def forward(self, input):
        residual = input
        x = self.conv1(input)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.conv2(x)
        x = self.bn2(x)
        x = self.relu(x)
        if self.downsample:
            residual = self.downsample(residual)
        x += residual
        x = self.relu(x)
        return x


class ResNet(nn.Module):
    def __init__(self, block, layers, num_classes=100):
        super(ResNet, self).__init__()
        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3, bias=False)
        self.bn1 = nn.BatchNorm2d(64)
        self.relu = nn.ReLU(inplace=True)
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=0, ceil_mode=True)
        self.layer1 = self._make_layer(block, 64, 64, layers[0])
        self.layer2 = self._make_layer(block, 64, 128, layers[1], stride=2)
        self.layer3 = self._make_layer(block, 128, 256, layers[2], stride=2)
        self.layer4 = self._make_layer(block, 256, 512, layers[3], stride=2)

        self.avgpool = nn.AvgPool2d(7)
        self.fc = KAN([512, num_classes])

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
                m.weight.data.normal_(0, math.sqrt(2. / n))
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_()

    def _make_layer(self, block, in_channel, out_channel, num_block, stride=1):
        downsample = None
        if stride != 1 or in_channel != out_channel:
            downsample = nn.Sequential(
                nn.Conv2d(in_channel, out_channel, kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(out_channel),
            )

        layers = []
        layers.append(block(in_channel, out_channel, stride, downsample))
        for i in range(1, num_block):
            layers.append(block(out_channel, out_channel))

        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        x = self.avgpool(x)
        x = x.view(x.size(0), -1)
        x = self.fc(x)
        return x


# def resnet18(**kwargs):
#     model = ResNet(BasicBlock, [2, 2, 2, 2], **kwargs)
#     return model
#
# from torchinfo import  summary
#
# model = resnet18().cuda()
#
# summary(model, input_size=(1, 3, 224, 224), device='cuda')