Problem occurred~

[pandamax]

InvalidArgumentError (see above for traceback): No OpKernel was registered to support Op 'MaxPoolWithArgmax' with these attrs. Registered devices: [CPU], Registered kernels: device='GPU'; Targmax in [DT_INT64]; T in [DT_DOUBLE] device='GPU'; Targmax in [DT_INT64]; T in [DT_FLOAT] device='GPU'; Targmax in [DT_INT64]; T in [DT_HALF]

 [[Node: ENet_1/bottleneck1_0_main_max_pool = MaxPoolWithArgmax[T=DT_FLOAT, Targmax=DT_INT64, ksize=[1, 2, 2, 1], padding="SAME", strides=[1, 2, 2, 1]](ENet_1/initial_block_1_concat)]]

[pandamax]

python2.7 tensorflow_gpu 1.2

[kwotsin]

Which system are you running on and did you build TF from source? This seems to be a similar problem encountered in older versions of TF where MaxPoolWithArgmax and some other ops were not yet implemented. Please also try running with TF 1.3 to see if the problem persists.

[chandrakantkhandelwal]

I have used the code for training on some customised data and I was able to run it sucessfully with tensroflow-gpu 1.2 (code doosn't runs on CPU verison of tensorflow 1.2, as there are some ops not supported on CPU).

[kwotsin]

It seems like TF-CPU v1.2 does not have some ops supported, rather than it being an issue with the code. I suggest you to raise an issue on TF to verify this issue. I'll close this issue for now.

[gunshi]

H @kwotsin , I also get the same error:

InvalidArgumentError (see above for traceback): No OpKernel was registered to support Op 'MaxPoolWithArgmax' with these attrs. Registered devices: [CPU], Registered kernels:

However I'm running it with tensorflow 1.4(gpu support), on a machine with 2 gpus. From the message, it looks like the gpu device isn't getting registered, and there's no explicit statement like 'with device('/gpu/:0')' in the predict_segmentation.py code, how can I change it to correct this? Thanks!