train_dist

hi
I trained atten model and I used train_dist file for the training
when I want to test the model using the trained atten model some errors occurred:
can u help me?
untimeError: Error(s) in loading state_dict for ATTEN:
Missing key(s) in state_dict: "pretrained.layer1.0.conv2.weight", "pretrained.layer1.0.bn2.weight", "pretrained.layer1.0.bn2.bias", "pretrained.layer1.0.bn2.running_mean", "pretrained.layer1.0.bn2.running_var", "pretrained.layer1.0.downsample.0.weight", "pretrained.layer1.0.downsample.1.bias", "pretrained.layer1.0.downsample.1.running_mean", "pretrained.layer1.0.downsample.1.running_var", "pretrained.layer1.1.conv2.weight", "pretrained.layer1.1.bn2.weight", "pretrained.layer1.1.bn2.bias", "pretrained.layer1.1.bn2.running_mean", "pretrained.layer1.1.bn2.running_var", "pretrained.layer1.2.conv2.weight", "pretrained.layer1.2.bn2.weight", "pretrained.layer1.2.bn2.bias", "pretrained.layer1.2.bn2.running_mean", "pretrained.layer1.2.bn2.running_var", "pretrained.layer2.0.conv2.weight", "pretrained.layer2.0.bn2.weight", "pretrained.layer2.0.bn2.bias", "pretrained.layer2.0.bn2.running_mean", "pretrained.layer2.0.bn2.running_var", "pretrained.layer2.0.downsample.0.weight", "pretrained.layer2.0.downsample.1.bias", "pretrained.layer2.0.downsample.1.running_mean", "pretrained.layer2.0.downsample.1.running_var", "pretrained.layer2.1.conv2.weight", "pretrained.layer2.1.bn2.weight", "pretrained.layer2.1.bn2.bias", "pretrained.layer2.1.bn2.running_mean", "pretrained.layer2.1.bn2.running_var", "pretrained.layer2.2.conv2.weight", "pretrained.layer2.2.bn2.weight", "pretrained.layer2.2.bn2.bias", "pretrained.layer2.2.bn2.running_mean", "pretrained.layer2.2.bn2.running_var", "pretrained.layer2.3.conv2.weight", "pretrained.layer2.3.bn2.weight", "pretrained.layer2.3.bn2.bias", "pretrained.layer2.3.bn2.running_mean", "pretrained.layer2.3.bn2.running_var", "pretrained.layer3.0.conv2.weight", "pretrained.layer3.0.bn2.weight", "pretrained.layer3.0.bn2.bias", "pretrained.layer3.0.bn2.running_mean", "pretrained.layer3.0.bn2.running_var", "pretrained.layer3.0.downsample.0.weight", "pretrained.layer3.0.downsample.1.bias", "pretrained.layer3.0.downsample.1.running_mean", "pretrained.layer3.0.downsample.1.running_var", "pretrained.layer3.1.conv2.weight", "pretrained.layer3.1.bn2.weight", "pretrained.layer3.1.bn2.bias", "pretrained.layer3.1.bn2.running_mean", "pretrained.layer3.1.bn2.running_var", "pretrained.layer3.2.conv2.weight", "pretrained.layer3.2.bn2.weight", "pretrained.layer3.2.bn2.bias", "pretrained.layer3.2.bn2.running_mean", "pretrained.layer3.2.bn2.running_var", "pretrained.layer3.3.conv2.weight", "pretrained.layer3.3.bn2.weight", "pretrained.layer3.3.bn2.bias", "pretrained.layer3.3.bn2.running_mean", "pretrained.layer3.3.bn2.running_var", "pretrained.layer3.4.conv2.weight", "pretrained.layer3.4.bn2.weight", "pretrained.layer3.4.bn2.bias", "pretrained.layer3.4.bn2.running_mean", "pretrained.layer3.4.bn2.running_var", "pretrained.layer3.5.conv2.weight", "pretrained.layer3.5.bn2.weight", "pretrained.layer3.5.bn2.bias", "pretrained.layer3.5.bn2.running_mean", "pretrained.layer3.5.bn2.running_var", "pretrained.layer4.0.conv2.weight", "pretrained.layer4.0.bn2.weight", "pretrained.layer4.0.bn2.bias", "pretrained.layer4.0.bn2.running_mean", "pretrained.layer4.0.bn2.running_var", "pretrained.layer4.0.downsample.0.weight", "pretrained.layer4.0.downsample.1.bias", "pretrained.layer4.0.downsample.1.running_mean", "pretrained.layer4.0.downsample.1.running_var", "pretrained.layer4.1.conv2.weight", "pretrained.layer4.1.bn2.weight", "pretrained.layer4.1.bn2.bias", "pretrained.layer4.1.bn2.running_mean", "pretrained.layer4.1.bn2.running_var", "pretrained.layer4.2.conv2.weight", "pretrained.layer4.2.bn2.weight", "pretrained.layer4.2.bn2.bias", "pretrained.layer4.2.bn2.running_mean", "pretrained.layer4.2.bn2.running_var".
Unexpected key(s) in state_dict: "pretrained.layer1.0.conv2.conv.weight", "pretrained.layer1.0.conv2.bn0.weight", "pretrained.layer1.0.conv2.bn0.bias", "pretrained.layer1.0.conv2.bn0.running_mean", "pretrained.layer1.0.conv2.bn0.running_var", "pretrained.layer1.0.conv2.bn0.num_batches_tracked", "pretrained.layer1.0.conv2.fc1.weight", "pretrained.layer1.0.conv2.fc1.bias", "pretrained.layer1.0.conv2.bn1.weight", "pretrained.layer1.0.conv2.bn1.bias", "pretrained.layer1.0.conv2.bn1.running_mean", "pretrained.layer1.0.conv2.bn1.running_var", "pretrained.layer1.0.conv2.bn1.num_batches_tracked", "pretrained.layer1.0.conv2.fc2.weight", "pretrained.layer1.0.conv2.fc2.bias", "pretrained.layer1.0.downsample.2.weight", "pretrained.layer1.0.downsample.2.bias", "pretrained.layer1.0.downsample.2.running_mean", "pretrained.layer1.0.downsample.2.running_var", "pretrained.layer1.0.downsample.2.num_batches_tracked", "pretrained.layer1.1.conv2.conv.weight", "pretrained.layer1.1.conv2.bn0.weight", "pretrained.layer1.1.conv2.bn0.bias", "pretrained.layer1.1.conv2.bn0.running_mean", "pretrained.layer1.1.conv2.bn0.running_var", "pretrained.layer1.1.conv2.bn0.num_batches_tracked", "pretrained.layer1.1.conv2.fc1.weight", "pretrained.layer1.1.conv2.fc1.bias", "pretrained.layer1.1.conv2.bn1.weight", "pretrained.layer1.1.conv2.bn1.bias", "pretrained.layer1.1.conv2.bn1.running_mean", "pretrained.layer1.1.conv2.bn1.running_var", "pretrained.layer1.1.conv2.bn1.num_batches_tracked", "pretrained.layer1.1.conv2.fc2.weight", "pretrained.layer1.1.conv2.fc2.bias", "pretrained.layer1.2.conv2.conv.weight", "pretrained.layer1.2.conv2.bn0.weight", "pretrained.layer1.2.conv2.bn0.bias", "pretrained.layer1.2.conv2.bn0.running_mean", "pretrained.layer1.2.conv2.bn0.running_var", "pretrained.layer1.2.conv2.bn0.num_batches_tracked", "pretrained.layer1.2.conv2.fc1.weight", "pretrained.layer1.2.conv2.fc1.bias", "pretrained.layer1.2.conv2.bn1.weight", "pretrained.layer1.2.conv2.bn1.bias", "pretrained.layer1.2.conv2.bn1.running_mean", "pretrained.layer1.2.conv2.bn1.running_var", "pretrained.layer1.2.conv2.bn1.num_batches_tracked", "pretrained.layer1.2.conv2.fc2.weight", "pretrained.layer1.2.conv2.fc2.bias", "pretrained.layer2.0.conv2.conv.weight", "pretrained.layer2.0.conv2.bn0.weight", "pretrained.layer2.0.conv2.bn0.bias", "pretrained.layer2.0.conv2.bn0.running_mean", "pretrained.layer2.0.conv2.bn0.running_var", "pretrained.layer2.0.conv2.bn0.num_batches_tracked", "pretrained.layer2.0.conv2.fc1.weight", "pretrained.layer2.0.conv2.fc1.bias", "pretrained.layer2.0.conv2.bn1.weight", "pretrained.layer2.0.conv2.bn1.bias", "pretrained.layer2.0.conv2.bn1.running_mean", "pretrained.layer2.0.conv2.bn1.running_var", "pretrained.layer2.0.conv2.bn1.num_batches_tracked", "pretrained.layer2.0.conv2.fc2.weight", "pretrained.layer2.0.conv2.fc2.bias", "pretrained.layer2.0.downsample.2.weight", "pretrained.layer2.0.downsample.2.bias", "pretrained.layer2.0.downsample.2.running_mean", "pretrained.layer2.0.downsample.2.running_var", "pretrained.layer2.0.downsample.2.num_batches_tracked", "pretrained.layer2.1.conv2.conv.weight", "pretrained.layer2.1.conv2.bn0.weight", "pretrained.layer2.1.conv2.bn0.bias", "pretrained.layer2.1.conv2.bn0.running_mean", "pretrained.layer2.1.conv2.bn0.running_var", "pretrained.layer2.1.conv2.bn0.num_batches_tracked", "pretrained.layer2.1.conv2.fc1.weight", "pretrained.layer2.1.conv2.fc1.bias", "pretrained.layer2.1.conv2.bn1.weight", "pretrained.layer2.1.conv2.bn1.bias", "pretrained.layer2.1.conv2.bn1.running_mean", "pretrained.layer2.1.conv2.bn1.running_var", "pretrained.layer2.1.conv2.bn1.num_batches_tracked", "pretrained.layer2.1.conv2.fc2.weight", "pretrained.layer2.1.conv2.fc2.bias", "pretrained.layer2.2.conv2.conv.weight", "pretrained.layer2.2.conv2.bn0.weight", "pretrained.layer2.2.conv2.bn0.bias", "pretrained.layer2.2.conv2.bn0.running_mean", "pretrained.layer2.2.conv2.bn0.running_var", "pretrained.layer2.2.conv2.bn0.num_batches_tracked", "pretrained.layer2.2.conv2.fc1.weight", "pretrained.layer2.2.conv2.fc1.bias", "pretrained.layer2.2.conv2.bn1.weight", "pretrained.layer2.2.conv2.bn1.bias", "pretrained.layer2.2.conv2.bn1.running_mean", "pretrained.layer2.2.conv2.bn1.running_var", "pretrained.layer2.2.conv2.bn1.num_batches_tracked", "pretrained.layer2.2.conv2.fc2.weight", "pretrained.layer2.2.conv2.fc2.bias", "pretrained.layer2.3.conv2.conv.weight", "pretrained.layer2.3.conv2.bn0.weight", "pretrained.layer2.3.conv2.bn0.bias", "pretrained.layer2.3.conv2.bn0.running_mean", "pretrained.layer2.3.conv2.bn0.running_var", "pretrained.layer2.3.conv2.bn0.num_batches_tracked", "pretrained.layer2.3.conv2.fc1.weight", "pretrained.layer2.3.conv2.fc1.bias", "pretrained.layer2.3.conv2.bn1.weight", "pretrained.layer2.3.conv2.bn1.bias", "pretrained.layer2.3.conv2.bn1.running_mean", "pretrained.layer2.3.conv2.bn1.running_var", "pretrained.layer2.3.conv2.bn1.num_batches_tracked", "pretrained.layer2.3.conv2.fc2.weight", "pretrained.layer2.3.conv2.fc2.bias", "pretrained.layer3.0.conv2.conv.weight", "pretrained.layer3.0.conv2.bn0.weight", "pretrained.layer3.0.conv2.bn0.bias", "pretrained.layer3.0.conv2.bn0.running_mean", "pretrained.layer3.0.conv2.bn0.running_var", "pretrained.layer3.0.conv2.bn0.num_batches_tracked", "pretrained.layer3.0.conv2.fc1.weight", "pretrained.layer3.0.conv2.fc1.bias", "pretrained.layer3.0.conv2.bn1.weight", "pretrained.layer3.0.conv2.bn1.bias", "pretrained.layer3.0.conv2.bn1.running_mean", "pretrained.layer3.0.conv2.bn1.running_var", "pretrained.layer3.0.conv2.bn1.num_batches_tracked", "pretrained.layer3.0.conv2.fc2.weight", "pretrained.layer3.0.conv2.fc2.bias", "pretrained.layer3.0.downsample.2.weight", "pretrained.layer3.0.downsample.2.bias", "pretrained.layer3.0.downsample.2.running_mean", "pretrained.layer3.0.downsample.2.running_var", "pretrained.layer3.0.downsample.2.num_batches_tracked", "pretrained.layer3.1.conv2.conv.weight", "pretrained.layer3.1.conv2.bn0.weight", "pretrained.layer3.1.conv2.bn0.bias", "pretrained.layer3.1.conv2.bn0.running_mean", "pretrained.layer3.1.conv2.bn0.running_var", "pretrained.layer3.1.conv2.bn0.num_batches_tracked", "pretrained.layer3.1.conv2.fc1.weight", "pretrained.layer3.1.conv2.fc1.bias", "pretrained.layer3.1.conv2.bn1.weight", "pretrained.layer3.1.conv2.bn1.bias", "pretrained.layer3.1.conv2.bn1.running_mean", "pretrained.layer3.1.conv2.bn1.running_var", "pretrained.layer3.1.conv2.bn1.num_batches_tracked", "pretrained.layer3.1.conv2.fc2.weight", "pretrained.layer3.1.conv2.fc2.bias", "pretrained.layer3.2.conv2.conv.weight", "pretrained.layer3.2.conv2.bn0.weight", "pretrained.layer3.2.conv2.bn0.bias", "pretrained.layer3.2.conv2.bn0.running_mean", "pretrained.layer3.2.conv2.bn0.running_var", "pretrained.layer3.2.conv2.bn0.num_batches_tracked", "pretrained.layer3.2.conv2.fc1.weight", "pretrained.layer3.2.conv2.fc1.bias", "pretrained.layer3.2.conv2.bn1.weight", "pretrained.layer3.2.conv2.bn1.bias", "pretrained.layer3.2.conv2.bn1.running_mean", "pretrained.layer3.2.conv2.bn1.running_var", "pretrained.layer3.2.conv2.bn1.num_batches_tracked", "pretrained.layer3.2.conv2.fc2.weight", "pretrained.layer3.2.conv2.fc2.bias", "pretrained.layer3.3.conv2.conv.weight", "pretrained.layer3.3.conv2.bn0.weight", "pretrained.layer3.3.conv2.bn0.bias", "pretrained.layer3.3.conv2.bn0.running_mean", "pretrained.layer3.3.conv2.bn0.running_var", "pretrained.layer3.3.conv2.bn0.num_batches_tracked", "pretrained.layer3.3.conv2.fc1.weight", "pretrained.layer3.3.conv2.fc1.bias", "pretrained.layer3.3.conv2.bn1.weight", "pretrained.layer3.3.conv2.bn1.bias", "pretrained.layer3.3.conv2.bn1.running_mean", "pretrained.layer3.3.conv2.bn1.running_var", "pretrained.layer3.3.conv2.bn1.num_batches_tracked", "pretrained.layer3.3.conv2.fc2.weight", "pretrained.layer3.3.conv2.fc2.bias", "pretrained.layer3.4.conv2.conv.weight", "pretrained.layer3.4.conv2.bn0.weight", "pretrained.layer3.4.conv2.bn0.bias", "pretrained.layer3.4.conv2.bn0.running_mean", "pretrained.layer3.4.conv2.bn0.running_var", "pretrained.layer3.4.conv2.bn0.num_batches_tracked", "pretrained.layer3.4.conv2.fc1.weight", "pretrained.layer3.4.conv2.fc1.bias", "pretrained.layer3.4.conv2.bn1.weight", "pretrained.layer3.4.conv2.bn1.bias", "pretrained.layer3.4.conv2.bn1.running_mean", "pretrained.layer3.4.conv2.bn1.running_var", "pretrained.layer3.4.conv2.bn1.num_batches_tracked", "pretrained.layer3.4.conv2.fc2.weight", "pretrained.layer3.4.conv2.fc2.bias", "pretrained.layer3.5.conv2.conv.weight", "pretrained.layer3.5.conv2.bn0.weight", "pretrained.layer3.5.conv2.bn0.bias", "pretrained.layer3.5.conv2.bn0.running_mean", "pretrained.layer3.5.conv2.bn0.running_var", "pretrained.layer3.5.conv2.bn0.num_batches_tracked", "pretrained.layer3.5.conv2.fc1.weight", "pretrained.layer3.5.conv2.fc1.bias", "pretrained.layer3.5.conv2.bn1.weight", "pretrained.layer3.5.conv2.bn1.bias", "pretrained.layer3.5.conv2.bn1.running_mean", "pretrained.layer3.5.conv2.bn1.running_var", "pretrained.layer3.5.conv2.bn1.num_batches_tracked", "pretrained.layer3.5.conv2.fc2.weight", "pretrained.layer3.5.conv2.fc2.bias", "pretrained.layer4.0.conv2.conv.weight", "pretrained.layer4.0.conv2.bn0.weight", "pretrained.layer4.0.conv2.bn0.bias", "pretrained.layer4.0.conv2.bn0.running_mean", "pretrained.layer4.0.conv2.bn0.running_var", "pretrained.layer4.0.conv2.bn0.num_batches_tracked", "pretrained.layer4.0.conv2.fc1.weight", "pretrained.layer4.0.conv2.fc1.bias", "pretrained.layer4.0.conv2.bn1.weight", "pretrained.layer4.0.conv2.bn1.bias", "pretrained.layer4.0.conv2.bn1.running_mean", "pretrained.layer4.0.conv2.bn1.running_var", "pretrained.layer4.0.conv2.bn1.num_batches_tracked", "pretrained.layer4.0.conv2.fc2.weight", "pretrained.layer4.0.conv2.fc2.bias", "pretrained.layer4.0.downsample.2.weight", "pretrained.layer4.0.downsample.2.bias", "pretrained.layer4.0.downsample.2.running_mean", "pretrained.layer4.0.downsample.2.running_var", "pretrained.layer4.0.downsample.2.num_batches_tracked", "pretrained.layer4.1.conv2.conv.weight", "pretrained.layer4.1.conv2.bn0.weight", "pretrained.layer4.1.conv2.bn0.bias", "pretrained.layer4.1.conv2.bn0.running_mean", "pretrained.layer4.1.conv2.bn0.running_var", "pretrained.layer4.1.conv2.bn0.num_batches_tracked", "pretrained.layer4.1.conv2.fc1.weight", "pretrained.layer4.1.conv2.fc1.bias", "pretrained.layer4.1.conv2.bn1.weight", "pretrained.layer4.1.conv2.bn1.bias", "pretrained.layer4.1.conv2.bn1.running_mean", "pretrained.layer4.1.conv2.bn1.running_var", "pretrained.layer4.1.conv2.bn1.num_batches_tracked", "pretrained.layer4.1.conv2.fc2.weight", "pretrained.layer4.1.conv2.fc2.bias", "pretrained.layer4.2.conv2.conv.weight", "pretrained.layer4.2.conv2.bn0.weight", "pretrained.layer4.2.conv2.bn0.bias", "pretrained.layer4.2.conv2.bn0.running_mean", "pretrained.layer4.2.conv2.bn0.running_var", "pretrained.layer4.2.conv2.bn0.num_batches_tracked", "pretrained.layer4.2.conv2.fc1.weight", "pretrained.layer4.2.conv2.fc1.bias", "pretrained.layer4.2.conv2.bn1.weight", "pretrained.layer4.2.conv2.bn1.bias", "pretrained.layer4.2.conv2.bn1.running_mean", "pretrained.layer4.2.conv2.bn1.running_var", "pretrained.layer4.2.conv2.bn1.num_batches_tracked", "pretrained.layer4.2.conv2.fc2.weight", "pretrained.layer4.2.conv2.fc2.bias".
size mismatch for pretrained.conv1.0.weight: copying a param with shape torch.Size([32, 3, 3, 3]) from checkpoint, the shape in current model is torch.Size([64, 3, 3, 3]).
size mismatch for pretrained.conv1.1.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.1.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.1.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.1.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.3.weight: copying a param with shape torch.Size([32, 32, 3, 3]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for pretrained.conv1.4.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.4.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.4.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.4.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for pretrained.conv1.6.weight: copying a param with shape torch.Size([64, 32, 3, 3]) from checkpoint, the shape in current model is torch.Size([128, 64, 3, 3]).
size mismatch for pretrained.bn1.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for pretrained.bn1.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for pretrained.bn1.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for pretrained.bn1.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for pretrained.layer1.0.conv1.weight: copying a param with shape torch.Size([64, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 128, 1, 1]).
size mismatch for pretrained.layer1.0.downsample.1.weight: copying a param with shape torch.Size([256, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for pretrained.layer2.0.downsample.1.weight: copying a param with shape torch.Size([512, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for pretrained.layer3.0.downsample.1.weight: copying a param with shape torch.Size([1024, 512, 1, 1]) from checkpoint, the shape in current model is torch.Size([1024]).
size mismatch for pretrained.layer4.0.downsample.1.weight: copying a param with shape torch.Size([2048, 1024, 1, 1]) from checkpoint, the shape in current model is torch.Size([2048]).

Start with the train_dist entry point and the ATTEN model-loading path, then reproduce the reported state_dict error using the same training and testing flow. Compare the checkpoint keys with the keys expected by ATTEN; done means a model trained with train_dist can be loaded for testing without missing or unexpected keys.