OS: Ubuntu 16.04
SNPE version: 1.15.0 and 1.16.2
When I execute the following command:
cd snpe-1.15.0/bin/x86_64-linux-clang
./snpe-tensorflow-to-dlc --graph mobilenet_v2_1.0_inf_graph_freeze.pb -i "input" 224,224,3 --out_node=MobilenetV2/Predictions
It fails. The log is shown as following:
Traceback (most recent call last):
File "./snpe-tensorflow-to-dlc", line 122, in main
converter.convert(args.dlc, args.model_version, converter_command)
File "/root/snpe/snpe-1.15.0/bin/x86_64-linux-clang/converters/tensorflow/converter.py", line 304, in convert
self._convert_layers()
File "/root/snpe/snpe-1.15.0/bin/x86_64-linux-clang/converters/tensorflow/converter.py", line 340, in _convert_layers
descriptors = self._resolve_descriptors_from_nodes(graph_ops)
File "/root/snpe/snpe-1.15.0/bin/x86_64-linux-clang/converters/tensorflow/converter.py", line 439, in _resolve_descriptors_from_nodes
resolved_descriptors = resolver.resolve_layer(graph_matcher, self._graph_helper)
File "/root/snpe/snpe-1.15.0/bin/x86_64-linux-clang/converters/tensorflow/layers/eltwise.py", line 97, in resolve_layer
const_tensor = self._broadcast_tensor(const_tensor, eltwise_shape)File "/root/snpe/snpe-1.15.0/bin/x86_64-linux-clang/converters/tensorflow/layers/eltwise.py", line 127, in _broadcast_tensor
broadcasted_tensor += tensor
ValueError: non-broadcastable output operand with shape (14,14,64) doesn't match the broadcast shape (1,14,14,64)