Convert deep-head-pose-lite from pytorch to caffe fail

[yueyihua]

conv1.0 conv: blob1 conv1 was added to layers 545911448128:conv_blob1 was added to blobs add1 was added to layers 545911448416:add_blob1 was added to blobs WARNING: CANNOT FOUND blob 8775320 8775320:extra_blob1 was added to blobs conv1.1 batch_norm1 was added to layers 545911448272:batch_norm_blob1 was added to blobs bn_scale1 was added to layers conv1.2 relu1 was added to layers 545913981760:relu_blob1 was added to blobs maxpool max_pool1 was added to layers 545911447624:max_pool_blob1 was added to blobs WARNING: the output shape miss match at max_pool1: input torch.Size([1, 24, 112, 112]) output---Pytorch:torch.Size([1, 24, 56, 56])---Caffe:torch.Size([1, 24, 57, 57]) This is caused by the different implementation that ceil mode in caffe and the floor mode in pytorch. You can add the clip layer in caffe prototxt manually if shape mismatch error is caused in caffe. stage2.0.branch1.0 conv: max_pool_blob1 conv2 was added to layers 545911447768:conv_blob2 was added to blobs add2 was added to layers 545911418744:add_blob2 was added to blobs WARNING: CANNOT FOUND blob 545913982624 Traceback (most recent call last): File "hopenet_to_caffe.py", line 15, in pytorch_to_caffe.trans_net(net,input,name) File "./pytorch_to_caffe.py", line 786, in trans_net out = net.forward(input_var) File "./stable_hopenetlite.py", line 127, in forward x = self.stage2(x) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(*input, kwargs) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/container.py", line 117, in forward input = module(input) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(*input, *kwargs) File "./stable_hopenetlite.py", line 76, in forward out = torch.cat((self.branch1(x), self.branch2(x)), dim=1) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(input, kwargs) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/container.py", line 117, in forward input = module(input) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 722, in _call_impl result = self.forward(*input, **kwargs) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/batchnorm.py", line 113, in forward self.num_batches_tracked = self.num_batches_tracked + 1 File "./pytorch_to_caffe.py", line 532, in _add bottom=[log.blobs(input),log.blobs(args[0])], top=top_blobs) File "./Caffe/layer_param.py", line 33, in init self.bottom.extend(bottom) TypeError: None has type NoneType, but expected one of: bytes, unicode

[XiaoLaoDi]

@yueyihua have you solved your problem ? how did you solve it ?