Hi,
I added new layer using custom head but i face error in loading dict.How do i initialize the weights of new layer
learn1 = create_cnn(
data1,
resnet501,
cut=-2,
split_on=_resnet_split,
loss_func=MixedLoss(),
custom_head=UnetBlock(2048,256,256,get_base(resnet501(True),-1)[6]),
#F.binary_cross_entropy_with_logits,
#FocalLoss(),
#F.binary_cross_entropy_with_logits,
path=path,
metrics=[acc_02,f1_scorestd], callback_fns= partial(GradientClipping, clip=1))
), Sequential( (0): Conv2d(1024, 1024, kernel_size=(1, 1), stride=(1, 1)) (1): ConvTranspose2d(2048, 1024, kernel_size=(2, 2), stride=(2, 2)) (2): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True) (3): Linear(in_features=2048, out_features=28, bias=True) )])>
Error(s) in loading state_dict for Sequential:
Missing key(s) in state_dict: “1.x_conv.weight”, “1.x_conv.bias”, “1.tr_conv.weight”, “1.tr_conv.bias”, “1.bn.weight”, “1.bn.bias”, “1.bn.running_mean”, “1.bn.running_var”, “1.fc.weight”, “1.fc.bias”.
Unexpected key(s) in state_dict: “1.2.weight”, “1.2.bias”, “1.2.running_mean”, “1.2.running_var”, “1.2.num_batches_tracked”, “1.4.weight”, “1.4.bias”, “1.6.weight”, “1.6.bias”, “1.6.running_mean”, “1.6.running_var”, “1.6.num_batches_tracked”, “1.8.weight”, “1.8.bias”.