Giters

pytorch / glow

Compiler for Neural Network hardware accelerators

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

[Bug] error on compiling an ONNX model

luyaor opened this issue · comments

commented

When compiling following ONNX model with Glow, the compiler fails with errors.
image

Error message:

F0828 08:00:55.572165  3922 Error.cpp:123] exitOnError(Error) got an unexpected ErrorValue: 
Error message: Node has no ints attribute with this name

Error return stack:
--------------------------------------------------------------------------------
/root/dev/glow/include/glow/Importer/ProtobufLoader.h:95
--------------------------------------------------------------------------------
/root/dev/glow/lib/Importer/ONNXModelLoader.cpp:2950
--------------------------------------------------------------------------------
/root/dev/glow/lib/Importer/ONNXModelLoader.cpp:6273
--------------------------------------------------------------------------------
/root/dev/glow/lib/Importer/ONNXModelLoader.cpp:6405
--------------------------------------------------------------------------------
/root/dev/glow/lib/Importer/ONNXModelLoader.cpp:6445
--------------------------------------------------------------------------------
/root/dev/glow/lib/Importer/ONNXModelLoader.cpp:6452
--------------------------------------------------------------------------------
*** Check failure stack trace: ***
./bin/model-compiler[0xcbde7f]
./bin/model-compiler[0xcbc2a2]
./bin/model-compiler[0xcbe288]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7f7083f9b980]
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0xc7)[0x7f7081a4ce87]
/lib/x86_64-linux-gnu/libc.so.6(abort+0x141)[0x7f7081a4e7f1]
/usr/lib/x86_64-linux-gnu/libglog.so.0(+0xa1f9)[0x7f70857bf1f9]
/usr/lib/x86_64-linux-gnu/libglog.so.0(+0xc0cd)[0x7f70857c10cd]
/usr/lib/x86_64-linux-gnu/libglog.so.0(_ZN6google10LogMessage9SendToLogEv+0x283)[0x7f70857c2f33]
/usr/lib/x86_64-linux-gnu/libglog.so.0(_ZN6google10LogMessage5FlushEv+0xb8)[0x7f70857c0c28]
/usr/lib/x86_64-linux-gnu/libglog.so.0(_ZN6google15LogMessageFatalD2Ev+0x9)[0x7f70857c3999]
./bin/model-compiler[0x2a0318a]
./bin/model-compiler[0xa7d71b]
./bin/model-compiler[0x4d010a]
./bin/model-compiler[0x521510]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7f7081a2fc87]
./bin/model-compiler[0x4cee5a]

The model file and run script are as follows:

g3.onnx.zip

./bin/model-compiler -model ./g3.onnx -emit-bundle ./mybundle --backend=CPU

commented

Hi I checked this issue and it caused by not supporting ONNX opset version 13 for Unsqueeze as well as for ReduceSum operators. Also, I faced with second issue because glow doesn't support multi axis input/attribute for ReduceSum and ReduceMean operators.
@luyaor can you please guide me what is necessary requirements for contribution, I am not sure who should assign me to specific issue or how to request assignment on issue.