UserWarning: positional arguments and argument \"destination\" are deprecated - Pytorch nn.modules.module.state_dict()

Question

I am trying to manage the checkpoints of my Pytorch model through torch.save():

Pytorch 1.12.0 and Python 3.7

torch.save({
        'epoch': epoch,
        'model_state_dict': model.state_dict(),
        'optimizer_state_dict': optimizer.state_dict()
      }, full_path)

But I am getting the following warning for model.state_dict():

/home/francesco/anaconda3/envs/env/lib/python3.7/site-packages/torch/nn/modules/module.py:1384: UserWarning: positional arguments and argument "destination" are deprecated. nn.Module.state_dict will not accept them in the future. Refer to https://pytorch.org/docs/master/generated/torch.nn.Module.html#torch.nn.Module.state_dict for details.

I had a look at the implementation of state_dict() here but I still don't get why I am getting the error since len(args) should be 0:

def state_dict(self, *args, destination=None, prefix='', keep_vars=False):
        warn_msg = []
        if len(args) > 0:
            warn_msg.append('positional arguments')
            if destination is None:
                destination = args[0]
            if len(args) > 1 and prefix == '':
                prefix = args[1]
            if len(args) > 2 and keep_vars is False:
                keep_vars = args[2]

        if destination is not None:
            warn_msg.append('argument "destination"')
        else:
            destination = OrderedDict()
            destination._metadata = OrderedDict()

        if warn_msg:
            # DeprecationWarning is ignored by default
            warnings.warn(
                " and ".join(warn_msg) + " are deprecated. nn.Module.state_dict will not accept them in the future. "
                "Refer to https://pytorch.org/docs/master/generated/torch.nn.Module.html#torch.nn.Module.state_dict for details.")

        return self._state_dict_impl(destination, prefix, keep_vars)

For the sake of completeness, here's the model:

import torch
import torch.nn as nn
import torch.nn.functional as F


class CNN(nn.Module):
    def __init__(self):
        super(CNN, self).__init__()

        self.conv1 = nn.Conv3d(in_channels=1, out_channels=32, kernel_size=3, stride=1, padding=1)
        self.pool1 = nn.MaxPool3d(kernel_size=2) 
        
        self.conv2 = nn.Conv3d(in_channels=32, out_channels=64, kernel_size=3, stride=1, padding=1)
        self.pool2 = nn.MaxPool3d(kernel_size=2) 

        self.dropout = nn.Dropout(0.5)
        
        self.fc1 = nn.Linear(16 * 16 * 16 * 64, 2)
        self.sig1 = nn.Sigmoid()

    def forward(self, x):
        
        x = F.relu(self.pool1(self.conv1(x))) 
        x = F.relu(self.pool2(self.conv2(x))) 
        x = x.view(-1, 16 * 16 * 16 * 64)

        x = self.dropout(x)
        
        x = self.sig1(self.fc1(x))

        return x

Anyone knows what I am missing? Thank you!

Answer 1

I have the same error, as a result it is not logging dict training logs. I'm training using PyTorch Lightning in DDP. I works on single GPU but gives this warming on multi-gpu system with DDP.