using tensorflow 1.6.0
keras 2.2.4
ncsdk 1
The model is from keras, and it is MobileNet, which is called using keras.application. (ps. I didn't make any changes to the model.)
I converted the model into a protobuf file and tried to covert it into graph using mvNCCompile, getting this error.
Is there any idea why this error happen?
Is it caused by the version of tensorflow? if so, which version is the best one for compiling to graph?
error:
InvalidArgumentError (see above for traceback): NodeDef mentions attr 'dilations' not in Op output:T; attr=T:type,allowed=[DT_HALF, DT_FLOAT];
attr=strides:list(int); attr=use_cudnn_on_gpu:bool,default=true; attr=padding:string,allowed=["SAME", "VALID"];
attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW"]>;
NodeDef: conv1/convolution = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], padding="VALID",
strides=[1, 2, 2, 1], use_cudnn_on_gpu=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"]
(conv1_pad/Pad, conv1/kernel).
(Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).
[[Node: conv1/convolution = Conv2D[T=DT_FLOAT, data_format="NHWC",
dilations=[1, 1, 1, 1], padding="VALID", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true,
_device="/job:localhost/replica:0/task:0/device:CPU:0"](conv1_pad/Pad, conv1/kernel)]]
