KeyError: "Unable to open object (object 'bn_adjust.num_batches_tracked' doesn't exist)"

[xiaotian3]

CLIENT: Creating process /home/p920/anaconda3/envs/pytorch/bin/python -c "import sys; sys.path.append('/home/p920/vot-toolkit/native/trax/python');sys.path.append('/home/p920/vot-toolkit/tracker/examples/python'); import python_siamvgg"
CLIENT: Working directory is /tmp/tp2e6fda82_3ac8_4f5c_874a_227ad99f0323
CLIENT: Starting process
CLIENT: Setting up TraX with standard streams connection
@@Trax:hello "trax.image=path" "trax.region=rectangle" "trax.version=1"
CLIENT: Tracker process ID: 19059
CLIENT: Connection with tracker established.
@@Trax:initialize "file:///home/p920/vot-toolkit/vot-2018/sequences/ants1/color/00000001.jpg" "128.7200,458.3600,28.1100,71.0500"
@@Trax:status "128.72,458.36,28.11,71.05"
Traceback (most recent call last):
File "", line 1, in
File "/home/p920/vot-toolkit/tracker/examples/python/python_siamvgg.py", line 20, in
tracker = SiamVGGTracker(imagefile,selection)
File "/home/p920/vot-toolkit/tracker/examples/python/src/siamvggtracker.py", line 46, in init
load_net(NET_PATH,self.siam)
File "/home/p920/vot-toolkit/tracker/examples/python/src/siamvggtracker.py", line 20, in load_net
param = torch.from_numpy(np.asarray(h5f[k]))
File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
File "/home/p920/anaconda3/envs/pytorch/lib/python2.7/site-packages/h5py/_hl/group.py", line 262, in getitem
oid = h5o.open(self.id, self._e(name), lapl=self._lapl)
File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
File "h5py/h5o.pyx", line 190, in h5py.h5o.open
KeyError: "Unable to open object (object 'bn_adjust.num_batches_tracked' doesn't exist)"

could you help me ?
my pytroch 0.40

[leeyeehoo]

So it's the pytorch version. I think PyTorch 5.0 will encounter these issue.

You can read this code:

def load_net(fname, net):
with h5py.File(fname, 'r') as h5f:
for k, v in net.state_dict().items():
try:
param = torch.from_numpy(np.asarray(h5f[k]))
except KeyError:
print "no value"
else: v.copy_(param)

[Totoro-wen]

hello,after above way following the problem. my pytorch 1.0

Traceback (most recent call last):
  File "/home/aibc/Wen/Siamese/Pytorch-SiamVGG/demo2.py", line 104, in <module>
    tracker = SiamVGGTracker(imagefile, region)
  File "/home/aibc/Wen/Siamese/Pytorch-SiamVGG/src/siamvggtracker.py", line 66, in __init__
    load_net(NET_PATH, self.siam)
  File "/home/aibc/Wen/Siamese/Pytorch-SiamVGG/src/siamvggtracker.py", line 34, in load_net
    v.copy_(param)
RuntimeError: expand(torch.FloatTensor{[1]}, size=[]): the number of sizes provided (0) must be greater or equal to the number of dimensions in the tensor (1)

can you help me?
Thanks.

[leeyeehoo]

I only use PyTorch 0.4.0 for the test. So I'm not fully understand your problem.

[Totoro-wen]

I only use PyTorch 0.4.0 for the test. So I'm not fully understand your problem.

I'm sorry I didn't make myself clear
After I replace

# def load_net(fname, net):
#     h5f = h5py.File(fname, mode='r')
#     for k, v in net.state_dict().items():
#         param = torch.from_numpy(np.asarray(h5f[k]))
#         v.copy_(param)

with

# def load_net(fname, net):
#     with h5py.File(fname, 'r') as h5f:
#         for k, v in net.state_dict().items():
#             try:
#                 param = torch.from_numpy(np.asarray(h5f[k]))
#             except KeyError:
#                 print("no value\n")
#             else: v.copy_(param)

, the following problem occurs.

Traceback (most recent call last):
File "python_siamvgg.py", line 27, in
tracker = SiamVGGTracker(imagefile, selection)
File "/home/aibc/Wen/Siamese/SiamVGG/vot2018submission/tracker/examples/python/src/siamvggtracker.py", line 71, in init
load_net(NET_PATH, self.siam)
File "/home/aibc/Wen/Siamese/SiamVGG/vot2018submission/tracker/examples/python/src/siamvggtracker.py", line 24, in load_net
param = torch.from_numpy(np.asarray(h5f[k]))
File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
File "/usr/local/lib/python3.5/dist-packages/h5py/_hl/group.py", line 177, in getitem
oid = h5o.open(self.id, self._e(name), lapl=self._lapl)
File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
File "h5py/h5o.pyx", line 190, in h5py.h5o.open
KeyError: "Unable to open object (object 'bn_adjust.num_batches_tracked' doesn't exist)"
Exception ignored in: <bound method VOT.del of <vot.VOT object at 0x7f2201883b00>>
Traceback (most recent call last):
File "/home/aibc/Wen/Siamese/SiamVGG/vot2018submission/tracker/examples/python/vot.py", line 189, in del
File "/home/aibc/Wen/Siamese/SiamVGG/vot2018submission/tracker/examples/python/vot.py", line 183, in quit
LookupError: unknown encoding: ascii

And my pytorch is 0.4.0 under anacoda.

[leeyeehoo]

If you use 0.4.0, you don't need to replace anything. If you use 1.0, you need to replace it.