Hi, tried converting a frozen tensorflow graph from https://www.github.com/ildoonet/tf-openpose and encountered the following error. Wasn't able to figure out what's wrong, have anyone encountered similar issues?
snpe-tensorflow-to-dlc --graph ~/qualcommTfpose/models/ildoonet.pb --input_dim "image" 320,176,3 --dlc ildoonet.dlc --out_node "Openpose/concat_stage7"
[[Node: MobilenetV1/Conv2d_1_depthwise/depthwise = DepthwiseConv2dNative[T=DT_FLOAT, data_format="NHWC", padding="SAME", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"](MobilenetV1/Conv2d_0/Relu, MobilenetV1/Conv2d_1_depthwise/depthwise_weights/read)]]2018-03-12 15:03:58,756 - 126 - ERROR - Encountered Error: list index out of rangeTraceback (most recent call last):File "/home/user/snpe-sdk/bin/x86_64-linux-clang/snpe-tensorflow-to-dlc", line 120, in mainconverter.convert(args.dlc, args.model_version, converter_command)File "/home/user/snpe-sdk/lib/python/converters/tensorflow/converter.py", line 303, in convertself._convert_layers()File "/home/user/snpe-sdk/lib/python/converters/tensorflow/converter.py", line 353, in _convert_layersself._create_layers(descriptors)File "/home/user/snpe-sdk/lib/python/converters/tensorflow/converter.py", line 389, in _create_layersself._create_layer(layer_builder, descriptor)File "/home/user/snpe-sdk/lib/python/converters/tensorflow/converter.py", line 461, in _create_layerlayer_builder.build_layer(self._context, descriptor, inputs, outputs)File "/home/user/snpe-sdk/lib/python/converters/tensorflow/layers/convolution.py", line 318, in build_layerdescriptor.dilationX])File "/home/user/snpe-sdk/lib/python/converters/tensorflow/layers/convolution.py", line 345, in calculate_padding_sizepad_y = max(((output_size[0] - 1) * strides[0] + filter_h - input_size[0]), 0)IndexError: list index out of range
Did you solve this problem?
I have same problem...
This is a bug when SNPE tries to merge two sequential 'concat' nodes.
Versions :
tensorflow-gpu 1.6
cuda 9.0 cudnn 7.1.2
SNPE 1.25.0
Bug Position :
in 'converter.py' line 422 , function 'transform_layer'
After processing the 1st 'concat' node in 'transform_layer', the second 'concat' node's 'child_ops' of the 1st concat was set to null , causing 'index out of range' error.
it is more than one year but noone fix this. I have problem with yolov3.
Hi, so have you sovled this problem?
I have the same problem now on converting yolov3, but I have no idea about this problem.
Can you give me any advice?
Thanks