I got this error when building onnx model(converted from pytorch):
Traceback (most recent call last):
File "east_model.py", line 227, in <module>
to_tvm()
File "east_model.py", line 219, in to_tvm
mod,params = relay.frontend.from_onnx(model,shape_dict)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/frontend/onnx.py", line 1868, in from_onnx
mod, params = g.from_onnx(graph, opset)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/frontend/onnx.py", line 1696, in from_onnx
op = self._convert_operator(op_name, inputs, attr, opset)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/frontend/onnx.py", line 1796, in _convert_operator
sym = convert_map[op_name](inputs, attrs, self._params)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/frontend/onnx.py", line 1423, in _impl_v11
scale = infer_value_simulated(inputs[2], params).asnumpy()
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/frontend/common.py", line 522, in infer_value_simulated
output_value = infer_value(input_val, params)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/frontend/common.py", line 496, in infer_value
graph, lib, params = tvm.relay.build(func, target="llvm", params=params)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/build_module.py", line 249, in build
graph_json, mod, params = bld_mod.build(func, target, target_host, params)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/relay/build_module.py", line 119, in build
self._build(func, target, target_host)
File "/home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/_ffi/_ctypes/packed_func.py", line 213, in __call__
raise get_last_ffi_error()
tvm._ffi.base.TVMError: Traceback (most recent call last):
[bt] (8) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(std::_Function_handler<void (tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*), tvm::relay::backend::RelayBuildModule::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#3}>::_M_invoke(std::_Any_data const&, tvm::runtime::TVMArgs&&, tvm::runtime::TVMRetValue*&&)+0x17) [0x7fc4a259c2b7]
[bt] (7) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::backend::RelayBuildModule::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#3}::operator()(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const+0x1b5) [0x7fc4a259c1c5]
[bt] (6) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::backend::RelayBuildModule::BuildRelay(tvm::relay::Function, std::unordered_map<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, tvm::runtime::NDArray, std::hash<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::equal_to<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, tvm::runtime::NDArray> > > const&)+0xa8c) [0x7fc4a259b54c]
[bt] (5) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(tvm::build(tvm::Map<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, tvm::Array<tvm::tir::LoweredFunc, void>, void, void> const&, tvm::Target const&, tvm::BuildConfig const&)+0x559) [0x7fc4a21d0ba9]
[bt] (4) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(tvm::build(tvm::Map<tvm::Target, tvm::Array<tvm::tir::LoweredFunc, void>, void, void> const&, tvm::Target const&, tvm::BuildConfig const&)+0x6a5) [0x7fc4a21cfc05]
[bt] (3) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(tvm::codegen::Build(tvm::Array<tvm::tir::LoweredFunc, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)+0x8a9) [0x7fc4a2234f19]
[bt] (2) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(+0x130bc3f) [0x7fc4a265bc3f]
[bt] (1) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(tvm::codegen::LLVMModuleNode::Init(tvm::Array<tvm::tir::LoweredFunc, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >)+0x5bd) [0x7fc4a265f98d]
[bt] (0) /home/dalalaa/.local/lib/python3.7/site-packages/tvm-0.7.dev0-py3.7-linux-x86_64.egg/tvm/libtvm.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x32) [0x7fc4a1e98342]
File "/home/dalalaa/library/tvm/src/target/llvm/llvm_module.cc", line 220
TVMError: LLVM module verification failed with the following errors:
Both operands to ICmp instruction are not of the same type!
%496 = icmp eq i32 %60, i64 0
My onnx graph is as follow:
graph(%input.1 : Float(1, 3, 256, 256),
%extractor.features.0.weight : Float(64, 3, 3, 3),
%extractor.features.0.bias : Float(64),
%extractor.features.1.weight : Float(64),
%extractor.features.1.bias : Float(64),
%extractor.features.1.running_mean : Float(64),
%extractor.features.1.running_var : Float(64),
%extractor.features.1.num_batches_tracked : Long(),
%extractor.features.3.weight : Float(64, 64, 3, 3),
%extractor.features.3.bias : Float(64),
%extractor.features.4.weight : Float(64),
%extractor.features.4.bias : Float(64),
%extractor.features.4.running_mean : Float(64),
%extractor.features.4.running_var : Float(64),
%extractor.features.4.num_batches_tracked : Long(),
%extractor.features.7.weight : Float(128, 64, 3, 3),
%extractor.features.7.bias : Float(128),
%extractor.features.8.weight : Float(128),
%extractor.features.8.bias : Float(128),
%extractor.features.8.running_mean : Float(128),
%extractor.features.8.running_var : Float(128),
%extractor.features.8.num_batches_tracked : Long(),
%extractor.features.10.weight : Float(128, 128, 3, 3),
%extractor.features.10.bias : Float(128),
%extractor.features.11.weight : Float(128),
%extractor.features.11.bias : Float(128),
%extractor.features.11.running_mean : Float(128),
%extractor.features.11.running_var : Float(128),
%extractor.features.11.num_batches_tracked : Long(),
%extractor.features.14.weight : Float(256, 128, 3, 3),
%extractor.features.14.bias : Float(256),
%extractor.features.15.weight : Float(256),
%extractor.features.15.bias : Float(256),
%extractor.features.15.running_mean : Float(256),
%extractor.features.15.running_var : Float(256),
%extractor.features.15.num_batches_tracked : Long(),
%extractor.features.17.weight : Float(256, 256, 3, 3),
%extractor.features.17.bias : Float(256),
%extractor.features.18.weight : Float(256),
%extractor.features.18.bias : Float(256),
%extractor.features.18.running_mean : Float(256),
%extractor.features.18.running_var : Float(256),
%extractor.features.18.num_batches_tracked : Long(),
%extractor.features.20.weight : Float(256, 256, 3, 3),
%extractor.features.20.bias : Float(256),
%extractor.features.21.weight : Float(256),
%extractor.features.21.bias : Float(256),
%extractor.features.21.running_mean : Float(256),
%extractor.features.21.running_var : Float(256),
%extractor.features.21.num_batches_tracked : Long(),
%extractor.features.24.weight : Float(512, 256, 3, 3),
%extractor.features.24.bias : Float(512),
%extractor.features.25.weight : Float(512),
%extractor.features.25.bias : Float(512),
%extractor.features.25.running_mean : Float(512),
%extractor.features.25.running_var : Float(512),
%extractor.features.25.num_batches_tracked : Long(),
%extractor.features.27.weight : Float(512, 512, 3, 3),
%extractor.features.27.bias : Float(512),
%extractor.features.28.weight : Float(512),
%extractor.features.28.bias : Float(512),
%extractor.features.28.running_mean : Float(512),
%extractor.features.28.running_var : Float(512),
%extractor.features.28.num_batches_tracked : Long(),
%extractor.features.30.weight : Float(512, 512, 3, 3),
%extractor.features.30.bias : Float(512),
%extractor.features.31.weight : Float(512),
%extractor.features.31.bias : Float(512),
%extractor.features.31.running_mean : Float(512),
%extractor.features.31.running_var : Float(512),
%extractor.features.31.num_batches_tracked : Long(),
%extractor.features.34.weight : Float(512, 512, 3, 3),
%extractor.features.34.bias : Float(512),
%extractor.features.35.weight : Float(512),
%extractor.features.35.bias : Float(512),
%extractor.features.35.running_mean : Float(512),
%extractor.features.35.running_var : Float(512),
%extractor.features.35.num_batches_tracked : Long(),
%extractor.features.37.weight : Float(512, 512, 3, 3),
%extractor.features.37.bias : Float(512),
%extractor.features.38.weight : Float(512),
%extractor.features.38.bias : Float(512),
%extractor.features.38.running_mean : Float(512),
%extractor.features.38.running_var : Float(512),
%extractor.features.38.num_batches_tracked : Long(),
%extractor.features.40.weight : Float(512, 512, 3, 3),
%extractor.features.40.bias : Float(512),
%extractor.features.41.weight : Float(512),
%extractor.features.41.bias : Float(512),
%extractor.features.41.running_mean : Float(512),
%extractor.features.41.running_var : Float(512),
%extractor.features.41.num_batches_tracked : Long(),
%merge.conv1.weight : Float(128, 1024, 1, 1),
%merge.conv1.bias : Float(128),
%merge.bn1.weight : Float(128),
%merge.bn1.bias : Float(128),
%merge.bn1.running_mean : Float(128),
%merge.bn1.running_var : Float(128),
%merge.bn1.num_batches_tracked : Long(),
%merge.conv2.weight : Float(128, 128, 3, 3),
%merge.conv2.bias : Float(128),
%merge.bn2.weight : Float(128),
%merge.bn2.bias : Float(128),
%merge.bn2.running_mean : Float(128),
%merge.bn2.running_var : Float(128),
%merge.bn2.num_batches_tracked : Long(),
%merge.conv3.weight : Float(64, 384, 1, 1),
%merge.conv3.bias : Float(64),
%merge.bn3.weight : Float(64),
%merge.bn3.bias : Float(64),
%merge.bn3.running_mean : Float(64),
%merge.bn3.running_var : Float(64),
%merge.bn3.num_batches_tracked : Long(),
%merge.conv4.weight : Float(64, 64, 3, 3),
%merge.conv4.bias : Float(64),
%merge.bn4.weight : Float(64),
%merge.bn4.bias : Float(64),
%merge.bn4.running_mean : Float(64),
%merge.bn4.running_var : Float(64),
%merge.bn4.num_batches_tracked : Long(),
%merge.conv5.weight : Float(32, 192, 1, 1),
%merge.conv5.bias : Float(32),
%merge.bn5.weight : Float(32),
%merge.bn5.bias : Float(32),
%merge.bn5.running_mean : Float(32),
%merge.bn5.running_var : Float(32),
%merge.bn5.num_batches_tracked : Long(),
%merge.conv6.weight : Float(32, 32, 3, 3),
%merge.conv6.bias : Float(32),
%merge.bn6.weight : Float(32),
%merge.bn6.bias : Float(32),
%merge.bn6.running_mean : Float(32),
%merge.bn6.running_var : Float(32),
%merge.bn6.num_batches_tracked : Long(),
%merge.conv7.weight : Float(32, 32, 3, 3),
%merge.conv7.bias : Float(32),
%merge.bn7.weight : Float(32),
%merge.bn7.bias : Float(32),
%merge.bn7.running_mean : Float(32),
%merge.bn7.running_var : Float(32),
%merge.bn7.num_batches_tracked : Long(),
%output.conv1.weight : Float(1, 32, 1, 1),
%output.conv1.bias : Float(1),
%output.conv2.weight : Float(4, 32, 1, 1),
%output.conv2.bias : Float(4),
%output.conv3.weight : Float(1, 32, 1, 1),
%output.conv3.bias : Float(1)):
%147 : Float(1, 64, 256, 256) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input.1, %extractor.features.0.weight, %extractor.features.0.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
...
...
%173 : Float(1, 512, 32, 32) = onnx::Relu(%172) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:912:0
%174 : Float(1, 512, 32, 32) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%173, %extractor.features.27.weight, %extractor.features.27.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%175 : Float(1, 512, 32, 32) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%174, %extractor.features.28.weight, %extractor.features.28.bias, %extractor.features.28.running_mean, %extractor.features.28.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%176 : Float(1, 512, 32, 32) = onnx::Relu(%175) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:912:0
%177 : Float(1, 512, 32, 32) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%176, %extractor.features.30.weight, %extractor.features.30.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%178 : Float(1, 512, 32, 32) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%177, %extractor.features.31.weight, %extractor.features.31.bias, %extractor.features.31.running_mean, %extractor.features.31.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%179 : Float(1, 512, 32, 32) = onnx::Relu(%178) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:912:0
%180 : Float(1, 512, 16, 16) = onnx::MaxPool[ceil_mode=0, kernel_shape=[2, 2], pads=[0, 0, 0, 0], strides=[2, 2]](%179) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:488:0
%181 : Float(1, 512, 16, 16) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%180, %extractor.features.34.weight, %extractor.features.34.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%182 : Float(1, 512, 16, 16) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%181, %extractor.features.35.weight, %extractor.features.35.bias, %extractor.features.35.running_mean, %extractor.features.35.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%183 : Float(1, 512, 16, 16) = onnx::Relu(%182) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:912:0
%184 : Float(1, 512, 16, 16) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%183, %extractor.features.37.weight, %extractor.features.37.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%185 : Float(1, 512, 16, 16) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%184, %extractor.features.38.weight, %extractor.features.38.bias, %extractor.features.38.running_mean, %extractor.features.38.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%186 : Float(1, 512, 16, 16) = onnx::Relu(%185) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:912:0
%187 : Float(1, 512, 16, 16) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%186, %extractor.features.40.weight, %extractor.features.40.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%188 : Float(1, 512, 16, 16) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%187, %extractor.features.41.weight, %extractor.features.41.bias, %extractor.features.41.running_mean, %extractor.features.41.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%189 : Float(1, 512, 16, 16) = onnx::Relu(%188) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:912:0
%190 : Float(1, 512, 8, 8) = onnx::MaxPool[ceil_mode=0, kernel_shape=[2, 2], pads=[0, 0, 0, 0], strides=[2, 2]](%189) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:488:0
%191 : Tensor = onnx::Shape(%190)
%192 : Tensor = onnx::Constant[value={2}]()
%193 : Long() = onnx::Gather[axis=0](%191, %192) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%194 : Float() = onnx::Cast[to=1](%193) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%195 : Float() = onnx::Constant[value={2}]()
%196 : Float() = onnx::Mul(%194, %195)
%197 : Float() = onnx::Cast[to=1](%196) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%198 : Float() = onnx::Floor(%197) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%199 : Tensor = onnx::Shape(%190)
%200 : Tensor = onnx::Constant[value={3}]()
%201 : Long() = onnx::Gather[axis=0](%199, %200) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%202 : Float() = onnx::Cast[to=1](%201) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%203 : Float() = onnx::Constant[value={2}]()
%204 : Float() = onnx::Mul(%202, %203)
%205 : Float() = onnx::Cast[to=1](%204) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%206 : Float() = onnx::Floor(%205) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%207 : Tensor = onnx::Unsqueeze[axes=[0]](%198)
%208 : Tensor = onnx::Unsqueeze[axes=[0]](%206)
%209 : Tensor = onnx::Concat[axis=0](%207, %208)
%210 : Tensor = onnx::Constant[value=[ CPUFloatType{0} ]]()
%211 : Tensor = onnx::Shape(%190)
%212 : Tensor = onnx::Constant[value={0}]()
%213 : Tensor = onnx::Constant[value={0}]()
%214 : Tensor = onnx::Constant[value={2}]()
%215 : Tensor = onnx::Slice(%211, %213, %214, %212)
%216 : Tensor = onnx::Cast[to=7](%209)
%217 : Tensor = onnx::Concat[axis=0](%215, %216)
%218 : Float(1, 512, 16, 16) = onnx::Resize[coordinate_transformation_mode="align_corners", cubic_coeff_a=-0.75, mode="linear", nearest_mode="floor"](%190, %210, %210, %217) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2530:0
%219 : Float(1, 1024, 16, 16) = onnx::Concat[axis=1](%218, %180) # east_model.py:121:0
%220 : Float(1, 128, 16, 16) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%219, %merge.conv1.weight, %merge.conv1.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%221 : Float(1, 128, 16, 16) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%220, %merge.bn1.weight, %merge.bn1.bias, %merge.bn1.running_mean, %merge.bn1.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%222 : Float(1, 128, 16, 16) = onnx::Relu(%221) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%223 : Float(1, 128, 16, 16) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%222, %merge.conv2.weight, %merge.conv2.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%224 : Float(1, 128, 16, 16) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%223, %merge.bn2.weight, %merge.bn2.bias, %merge.bn2.running_mean, %merge.bn2.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%225 : Float(1, 128, 16, 16) = onnx::Relu(%224) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%226 : Tensor = onnx::Shape(%225)
%227 : Tensor = onnx::Constant[value={2}]()
%228 : Long() = onnx::Gather[axis=0](%226, %227) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%229 : Float() = onnx::Cast[to=1](%228) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%230 : Float() = onnx::Constant[value={2}]()
%231 : Float() = onnx::Mul(%229, %230)
%232 : Float() = onnx::Cast[to=1](%231) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%233 : Float() = onnx::Floor(%232) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%234 : Tensor = onnx::Shape(%225)
%235 : Tensor = onnx::Constant[value={3}]()
%236 : Long() = onnx::Gather[axis=0](%234, %235) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%237 : Float() = onnx::Cast[to=1](%236) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%238 : Float() = onnx::Constant[value={2}]()
%239 : Float() = onnx::Mul(%237, %238)
%240 : Float() = onnx::Cast[to=1](%239) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%241 : Float() = onnx::Floor(%240) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%242 : Tensor = onnx::Unsqueeze[axes=[0]](%233)
%243 : Tensor = onnx::Unsqueeze[axes=[0]](%241)
%244 : Tensor = onnx::Concat[axis=0](%242, %243)
%245 : Tensor = onnx::Constant[value=[ CPUFloatType{0} ]]()
%246 : Tensor = onnx::Shape(%225)
%247 : Tensor = onnx::Constant[value={0}]()
%248 : Tensor = onnx::Constant[value={0}]()
%249 : Tensor = onnx::Constant[value={2}]()
%250 : Tensor = onnx::Slice(%246, %248, %249, %247)
%251 : Tensor = onnx::Cast[to=7](%244)
%252 : Tensor = onnx::Concat[axis=0](%250, %251)
%253 : Float(1, 128, 32, 32) = onnx::Resize[coordinate_transformation_mode="align_corners", cubic_coeff_a=-0.75, mode="linear", nearest_mode="floor"](%225, %245, %245, %252) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2530:0
%254 : Float(1, 384, 32, 32) = onnx::Concat[axis=1](%253, %170) # east_model.py:126:0
%255 : Float(1, 64, 32, 32) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%254, %merge.conv3.weight, %merge.conv3.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%256 : Float(1, 64, 32, 32) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%255, %merge.bn3.weight, %merge.bn3.bias, %merge.bn3.running_mean, %merge.bn3.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%257 : Float(1, 64, 32, 32) = onnx::Relu(%256) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%258 : Float(1, 64, 32, 32) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%257, %merge.conv4.weight, %merge.conv4.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%259 : Float(1, 64, 32, 32) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%258, %merge.bn4.weight, %merge.bn4.bias, %merge.bn4.running_mean, %merge.bn4.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%260 : Float(1, 64, 32, 32) = onnx::Relu(%259) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%261 : Tensor = onnx::Shape(%260)
%262 : Tensor = onnx::Constant[value={2}]()
%263 : Long() = onnx::Gather[axis=0](%261, %262) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%264 : Float() = onnx::Cast[to=1](%263) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%265 : Float() = onnx::Constant[value={2}]()
%266 : Float() = onnx::Mul(%264, %265)
%267 : Float() = onnx::Cast[to=1](%266) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%268 : Float() = onnx::Floor(%267) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%269 : Tensor = onnx::Shape(%260)
%270 : Tensor = onnx::Constant[value={3}]()
%271 : Long() = onnx::Gather[axis=0](%269, %270) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%272 : Float() = onnx::Cast[to=1](%271) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%273 : Float() = onnx::Constant[value={2}]()
%274 : Float() = onnx::Mul(%272, %273)
%275 : Float() = onnx::Cast[to=1](%274) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%276 : Float() = onnx::Floor(%275) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2493:0
%277 : Tensor = onnx::Unsqueeze[axes=[0]](%268)
%278 : Tensor = onnx::Unsqueeze[axes=[0]](%276)
%279 : Tensor = onnx::Concat[axis=0](%277, %278)
%280 : Tensor = onnx::Constant[value=[ CPUFloatType{0} ]]()
%281 : Tensor = onnx::Shape(%260)
%282 : Tensor = onnx::Constant[value={0}]()
%283 : Tensor = onnx::Constant[value={0}]()
%284 : Tensor = onnx::Constant[value={2}]()
%285 : Tensor = onnx::Slice(%281, %283, %284, %282)
%286 : Tensor = onnx::Cast[to=7](%279)
%287 : Tensor = onnx::Concat[axis=0](%285, %286)
%288 : Float(1, 64, 64, 64) = onnx::Resize[coordinate_transformation_mode="align_corners", cubic_coeff_a=-0.75, mode="linear", nearest_mode="floor"](%260, %280, %280, %287) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:2530:0
%289 : Float(1, 192, 64, 64) = onnx::Concat[axis=1](%288, %160) # east_model.py:131:0
%290 : Float(1, 32, 64, 64) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%289, %merge.conv5.weight, %merge.conv5.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%291 : Float(1, 32, 64, 64) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%290, %merge.bn5.weight, %merge.bn5.bias, %merge.bn5.running_mean, %merge.bn5.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%292 : Float(1, 32, 64, 64) = onnx::Relu(%291) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%293 : Float(1, 32, 64, 64) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%292, %merge.conv6.weight, %merge.conv6.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%294 : Float(1, 32, 64, 64) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%293, %merge.bn6.weight, %merge.bn6.bias, %merge.bn6.running_mean, %merge.bn6.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%295 : Float(1, 32, 64, 64) = onnx::Relu(%294) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%296 : Float(1, 32, 64, 64) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%295, %merge.conv7.weight, %merge.conv7.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%297 : Float(1, 32, 64, 64) = onnx::BatchNormalization[epsilon=1e-05, momentum=0.9](%296, %merge.bn7.weight, %merge.bn7.bias, %merge.bn7.running_mean, %merge.bn7.running_var) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:1670:0
%298 : Float(1, 32, 64, 64) = onnx::Relu(%297) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/functional.py:914:0
%299 : Float(1, 1, 64, 64) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%298, %output.conv1.weight, %output.conv1.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%300 : Float(1, 1, 64, 64) = onnx::Sigmoid(%299) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/activation.py:271:0
%301 : Float(1, 4, 64, 64) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%298, %output.conv2.weight, %output.conv2.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%302 : Float(1, 4, 64, 64) = onnx::Sigmoid(%301) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/activation.py:271:0
%303 : Float() = onnx::Constant[value={512}]()
%304 : Float(1, 4, 64, 64) = onnx::Mul(%302, %303)
%305 : Float(1, 1, 64, 64) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%298, %output.conv3.weight, %output.conv3.bias) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/conv.py:342:0
%306 : Float(1, 1, 64, 64) = onnx::Sigmoid(%305) # /home/dalalaa/miniconda3/envs/py37/lib/python3.7/site-packages/torch/nn/modules/activation.py:271:0
%307 : Float() = onnx::Constant[value={0.5}]()
%308 : Float(1, 1, 64, 64) = onnx::Sub(%306, %307)
%309 : Float() = onnx::Constant[value={3.14159}]()
%310 : Float(1, 1, 64, 64) = onnx::Mul(%308, %309)
%311 : Float(1, 5, 64, 64) = onnx::Concat[axis=1](%304, %310) # east_model.py:158:0
return (%300, %311)
I think this error is caused by the third onnx::Resize op in my graph. How could I solve it ?
Please tell me if more detail is needed. Looking forward to your reply.