The log is
```
Traceback (most recent call last):
File
"/home/data/git/OnlineActionRecognition/tools/model_convert/tsm_model_to_tvm/tvm_auto_tuning.py",
line 174, in <module>
tune_and_evaluate(tuning_option)
File
"/home/data/git/OnlineActionRecognition/tools/model_convert/tsm_model_to_tvm/tvm_auto_tuning.py",
line 124, in tune_and_evaluate
mod, params = get_tvm_module(torch_inputs[0])
File
"/home/data/git/OnlineActionRecognition/tools/model_convert/tsm_model_to_tvm/tvm_auto_tuning.py",
line 48, in get_tvm_module
relay_module, params = tvm.relay.frontend.from_onnx(onnx_model,
shape=shape_dict, dtype="float16")
File
"/home/cyh/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/relay/frontend/onnx.py",
line 1879, in from_onnx
mod, params = g.from_onnx(graph, opset)
File
"/home/cyh/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/relay/frontend/onnx.py",
line 1726, in from_onnx
return IRModule.from_expr(func), self._params
File
"/home/cyh/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/ir/module.py",
line 223, in from_expr
return _ffi_api.Module_FromExpr(expr, funcs, defs)
File "tvm/_ffi/_cython/./packed_func.pxi", line 308, in
tvm._ffi._cy3.core.PackedFuncBase.__call__
File "tvm/_ffi/_cython/./packed_func.pxi", line 243, in
tvm._ffi._cy3.core.FuncCall
File "tvm/_ffi/_cython/./packed_func.pxi", line 232, in
tvm._ffi._cy3.core.FuncCall3
File "tvm/_ffi/_cython/./base.pxi", line 159, in tvm._ffi._cy3.core.CALL
tvm._ffi.base.TVMError: Traceback (most recent call last):
[bt] (8)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(TVMFuncCall+0x61)
[0x7fc46241ff61]
[bt] (7)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(+0x4ec44c)
[0x7fc461cba44c]
[bt] (6)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::IRModule::FromExpr(tvm::RelayExpr
const&, tvm::Map<tvm::GlobalVar, tvm::BaseFunc, void, void> const&,
tvm::Map<tvm::GlobalTypeVar, tvm::TypeData, void, void> const&)+0x189)
[0x7fc461cb7739]
[bt] (5)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::IRModuleNode::Add(tvm::GlobalVar
const&, tvm::BaseFunc const&, bool)+0xdc) [0x7fc461cb70ec]
[bt] (4)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::RunTypeCheck(tvm::IRModule
const&, tvm::GlobalVar const&, tvm::relay::Function)+0x277) [0x7fc461cb6377]
[bt] (3)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::InferType(tvm::relay::Function
const&, tvm::IRModule const&, tvm::GlobalVar const&)+0x1d4) [0x7fc4622b0c64]
[bt] (2)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::TypeInferencer::Infer(tvm::RelayExpr)+0x86)
[0x7fc4622b0486]
[bt] (1)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::ErrorReporter::RenderErrors(tvm::IRModule
const&, bool)+0x2816) [0x7fc461ca8ee6]
[bt] (0)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x32)
[0x7fc461bac5c2]
[bt] (8)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(+0x4ec44c)
[0x7fc461cba44c]
[bt] (7)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::IRModule::FromExpr(tvm::RelayExpr
const&, tvm::Map<tvm::GlobalVar, tvm::BaseFunc, void, void> const&,
tvm::Map<tvm::GlobalTypeVar, tvm::TypeData, void, void> const&)+0x189)
[0x7fc461cb7739]
[bt] (6)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::IRModuleNode::Add(tvm::GlobalVar
const&, tvm::BaseFunc const&, bool)+0xdc) [0x7fc461cb70ec]
[bt] (5)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::RunTypeCheck(tvm::IRModule
const&, tvm::GlobalVar const&, tvm::relay::Function)+0x277) [0x7fc461cb6377]
[bt] (4)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::InferType(tvm::relay::Function
const&, tvm::IRModule const&, tvm::GlobalVar const&)+0x1d4) [0x7fc4622b0c64]
[bt] (3)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::TypeInferencer::Infer(tvm::RelayExpr)+0x55)
[0x7fc4622b0455]
[bt] (2)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::TypeSolver::Solve()+0x3b0)
[0x7fc4621b0f40]
[bt] (1)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(std::_Function_handler<void
(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*),
tvm::runtime::TypedPackedFunc<bool (tvm::Array<tvm::Type, void> const&, int,
tvm::Attrs const&, tvm::TypeReporter const&)>::AssignTypedLambda<bool
(*)(tvm::Array<tvm::Type, void> const&, int, tvm::Attrs const&,
tvm::TypeReporter const&)>(bool (*)(tvm::Array<tvm::Type, void> const&, int,
tvm::Attrs const&, tvm::TypeReporter const&))::{lambda(tvm::runtime::TVMArgs
const&, tvm::runtime::TVMRetValue*)#1}>::_M_invoke(std::_Any_data const&,
tvm::runtime::TVMArgs&&, tvm::runtime::TVMRetValue*&&)+0xd4) [0x7fc461fad504]
[bt] (0)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::BroadcastRel(tvm::Array<tvm::Type,
void> const&, int, tvm::Attrs const&, tvm::TypeReporter const&)+0xb4d)
[0x7fc46215641d]
[bt] (8)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(+0x4ec44c)
[0x7fc461cba44c]
[bt] (7)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::IRModule::FromExpr(tvm::RelayExpr
const&, tvm::Map<tvm::GlobalVar, tvm::BaseFunc, void, void> const&,
tvm::Map<tvm::GlobalTypeVar, tvm::TypeData, void, void> const&)+0x189)
[0x7fc461cb7739]
[bt] (6)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::IRModuleNode::Add(tvm::GlobalVar
const&, tvm::BaseFunc const&, bool)+0xdc) [0x7fc461cb70ec]
[bt] (5)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::RunTypeCheck(tvm::IRModule
const&, tvm::GlobalVar const&, tvm::relay::Function)+0x277) [0x7fc461cb6377]
[bt] (4)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::InferType(tvm::relay::Function
const&, tvm::IRModule const&, tvm::GlobalVar const&)+0x1d4) [0x7fc4622b0c64]
[bt] (3)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::TypeInferencer::Infer(tvm::RelayExpr)+0x55)
[0x7fc4622b0455]
[bt] (2)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::TypeSolver::Solve()+0x3b0)
[0x7fc4621b0f40]
[bt] (1)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(std::_Function_handler<void
(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*),
tvm::runtime::TypedPackedFunc<bool (tvm::Array<tvm::Type, void> const&, int,
tvm::Attrs const&, tvm::TypeReporter const&)>::AssignTypedLambda<bool
(*)(tvm::Array<tvm::Type, void> const&, int, tvm::Attrs const&,
tvm::TypeReporter const&)>(bool (*)(tvm::Array<tvm::Type, void> const&, int,
tvm::Attrs const&, tvm::TypeReporter const&))::{lambda(tvm::runtime::TVMArgs
const&, tvm::runtime::TVMRetValue*)#1}>::_M_invoke(std::_Any_data const&,
tvm::runtime::TVMArgs&&, tvm::runtime::TVMRetValue*&&)+0xd4) [0x7fc461fad504]
[bt] (0)
/home/data/cyh_home/miniconda3/envs/py36-pytorch/lib/python3.6/site-packages/tvm-0.7.dev1-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::relay::BroadcastRel(tvm::Array<tvm::Type,
void> const&, int, tvm::Attrs const&, tvm::TypeReporter const&)+0xb4d)
[0x7fc46215641d]
File "/home/data/git/tvm/src/ir/error.cc", line 133
TVMError:
Error(s) have occurred. The program has been annotated with them:
In `main`:
v0.0.4
fn (%input: Tensor[(1, 24, 224, 224), float16], %base_model.conv1.weight:
Tensor[(64, 3, 7, 7), float16], %base_model.bn1.weight: Tensor[(64), float16],
%base_model.bn1.bias: Tensor[(64), float16], %base_model.bn1.running_mean:
Tensor[(64), float16], %base_model.bn1.running_var: Tensor[(64), float16],
%v344: Tensor[(1, 1, 8, 56, 56), float16], %v346: Tensor[(1, 1, 8, 56, 56),
float16], %base_model.layer1.0.conv1.net.weight: Tensor[(64, 64, 1, 1),
float16], %base_model.layer1.0.bn1.weight: Tensor[(64), float16],
%base_model.layer1.0.bn1.bias: Tensor[(64), float16],
%base_model.layer1.0.bn1.running_mean: Tensor[(64), float16],
%base_model.layer1.0.bn1.running_var: Tensor[(64), float16],
%base_model.layer1.0.conv2.weight: Tensor[(64, 64, 3, 3), float16],
%base_model.layer1.0.bn2.weight: Tensor[(64), float16],
%base_model.layer1.0.bn2.bias: Tensor[(64), float16],
%base_model.layer1.0.bn2.running_mean: Tensor[(64), float16],
%base_model.layer1.0.bn2.running_var: Tensor[(64), float16],
%base_model.layer1.0.conv3.weight: Tensor[(256, 64, 1, 1), float16],
%base_model.layer1.0.bn3.weight: Tensor[(256), float16],
%base_model.layer1.0.bn3.bias: Tensor[(256), float16],
%base_model.layer1.0.bn3.running_mean: Tensor[(256), float16],
%base_model.layer1.0.bn3.running_var: Tensor[(256), float16],
%base_model.layer1.0.downsample.0.weight: Tensor[(256, 64, 1, 1), float16],
%base_model.layer1.0.downsample.1.weight: Tensor[(256), float16],
%base_model.layer1.0.downsample.1.bias: Tensor[(256), float16],
%base_model.layer1.0.downsample.1.running_mean: Tensor[(256), float16],
%base_model.layer1.0.downsample.1.running_var: Tensor[(256), float16], %v410:
Tensor[(1, 1, 32, 56, 56), float16], %v412: Tensor[(1, 1, 32, 56, 56),
float16], %base_model.layer1.1.conv1.net.weight: Tensor[(64, 256, 1, 1),
float16], %base_model.layer1.1.bn1.weight: Tensor[(64), float16],
%base_model.layer1.1.bn1.bias: Tensor[(64), float16],
%base_model.layer1.1.bn1.running_mean: Tensor[(64), float16],
%base_model.layer1.1.bn1.running_var: Tensor[(64), float16],
%base_model.layer1.1.conv2.weight: Tensor[(64, 64, 3, 3), float16],
%base_model.layer1.1.bn2.weight: Tensor[(64), float16],
%base_model.layer1.1.bn2.bias: Tensor[(64), float16],
%base_model.layer1.1.bn2.running_mean: Tensor[(64), float16],
%base_model.layer1.1.bn2.running_var: Tensor[(64), float16],
%base_model.layer1.1.conv3.weight: Tensor[(256, 64, 1, 1), float16],
%base_model.layer1.1.bn3.weight: Tensor[(256), float16],
%base_model.layer1.1.bn3.bias: Tensor[(256), float16],
%base_model.layer1.1.bn3.running_mean: Tensor[(256), float16],
%base_model.layer1.1.bn3.running_var: Tensor[(256), float16], %v474: Tensor[(1,
1, 32, 56, 56), float16], %v476: Tensor[(1, 1, 32, 56, 56), float16],
%base_model.layer1.2.conv1.net.weight: Tensor[(64, 256, 1, 1), float16],
%base_model.layer1.2.bn1.weight: Tensor[(64), float16],
%base_model.layer1.2.bn1.bias: Tensor[(64), float16],
%base_model.layer1.2.bn1.running_mean: Tensor[(64), float16],
%base_model.layer1.2.bn1.running_var: Tensor[(64), float16],
%base_model.layer1.2.conv2.weight: Tensor[(64, 64, 3, 3), float16],
%base_model.layer1.2.bn2.weight: Tensor[(64), float16],
%base_model.layer1.2.bn2.bias: Tensor[(64), float16],
%base_model.layer1.2.bn2.running_mean: Tensor[(64), float16],
%base_model.layer1.2.bn2.running_var: Tensor[(64), float16],
%base_model.layer1.2.conv3.weight: Tensor[(256, 64, 1, 1), float16],
%base_model.layer1.2.bn3.weight: Tensor[(256), float16],
%base_model.layer1.2.bn3.bias: Tensor[(256), float16],
%base_model.layer1.2.bn3.running_mean: Tensor[(256), float16],
%base_model.layer1.2.bn3.running_var: Tensor[(256), float16], %v538: Tensor[(1,
1, 32, 56, 56), float16], %v540: Tensor[(1, 1, 32, 56, 56), float16],
%base_model.layer2.0.conv1.net.weight: Tensor[(128, 256, 1, 1), float16],
%base_model.layer2.0.bn1.weight: Tensor[(128), float16],
%base_model.layer2.0.bn1.bias: Tensor[(128), float16],
%base_model.layer2.0.bn1.running_mean: Tensor[(128), float16],
%base_model.layer2.0.bn1.running_var: Tensor[(128), float16],
%base_model.layer2.0.conv2.weight: Tensor[(128, 128, 3, 3), float16],
%base_model.layer2.0.bn2.weight: Tensor[(128), float16],
%base_model.layer2.0.bn2.bias: Tensor[(128), float16],
%base_model.layer2.0.bn2.running_mean: Tensor[(128), float16],
%base_model.layer2.0.bn2.running_var: Tensor[(128), float16],
%base_model.layer2.0.conv3.weight: Tensor[(512, 128, 1, 1), float16],
%base_model.layer2.0.bn3.weight: Tensor[(512), float16],
%base_model.layer2.0.bn3.bias: Tensor[(512), float16],
%base_model.layer2.0.bn3.running_mean: Tensor[(512), float16],
%base_model.layer2.0.bn3.running_var: Tensor[(512), float16],
%base_model.layer2.0.downsample.0.weight: Tensor[(512, 256, 1, 1), float16],
%base_model.layer2.0.downsample.1.weight: Tensor[(512), float16],
%base_model.layer2.0.downsample.1.bias: Tensor[(512), float16],
%base_model.layer2.0.downsample.1.running_mean: Tensor[(512), float16],
%base_model.layer2.0.downsample.1.running_var: Tensor[(512), float16], %v604:
Tensor[(1, 1, 64, 28, 28), float16], %v606: Tensor[(1, 1, 64, 28, 28),
float16], %base_model.layer2.1.conv1.net.weight: Tensor[(128, 512, 1, 1),
float16], %base_model.layer2.1.bn1.weight: Tensor[(128), float16],
%base_model.layer2.1.bn1.bias: Tensor[(128), float16],
%base_model.layer2.1.bn1.running_mean: Tensor[(128), float16],
%base_model.layer2.1.bn1.running_var: Tensor[(128), float16],
%base_model.layer2.1.conv2.weight: Tensor[(128, 128, 3, 3), float16],
%base_model.layer2.1.bn2.weight: Tensor[(128), float16],
%base_model.layer2.1.bn2.bias: Tensor[(128), float16],
%base_model.layer2.1.bn2.running_mean: Tensor[(128), float16],
%base_model.layer2.1.bn2.running_var: Tensor[(128), float16],
%base_model.layer2.1.conv3.weight: Tensor[(512, 128, 1, 1), float16],
%base_model.layer2.1.bn3.weight: Tensor[(512), float16],
%base_model.layer2.1.bn3.bias: Tensor[(512), float16],
%base_model.layer2.1.bn3.running_mean: Tensor[(512), float16],
%base_model.layer2.1.bn3.running_var: Tensor[(512), float16], %v668: Tensor[(1,
1, 64, 28, 28), float16], %v670: Tensor[(1, 1, 64, 28, 28), float16],
%base_model.layer2.2.conv1.net.weight: Tensor[(128, 512, 1, 1), float16],
%base_model.layer2.2.bn1.weight: Tensor[(128), float16],
%base_model.layer2.2.bn1.bias: Tensor[(128), float16],
%base_model.layer2.2.bn1.running_mean: Tensor[(128), float16],
%base_model.layer2.2.bn1.running_var: Tensor[(128), float16],
%base_model.layer2.2.conv2.weight: Tensor[(128, 128, 3, 3), float16],
%base_model.layer2.2.bn2.weight: Tensor[(128), float16],
%base_model.layer2.2.bn2.bias: Tensor[(128), float16],
%base_model.layer2.2.bn2.running_mean: Tensor[(128), float16],
%base_model.layer2.2.bn2.running_var: Tensor[(128), float16],
%base_model.layer2.2.conv3.weight: Tensor[(512, 128, 1, 1), float16],
%base_model.layer2.2.bn3.weight: Tensor[(512), float16],
%base_model.layer2.2.bn3.bias: Tensor[(512), float16],
%base_model.layer2.2.bn3.running_mean: Tensor[(512), float16],
%base_model.layer2.2.bn3.running_var: Tensor[(512), float16], %v732: Tensor[(1,
1, 64, 28, 28), float16], %v734: Tensor[(1, 1, 64, 28, 28), float16],
%base_model.layer2.3.conv1.net.weight: Tensor[(128, 512, 1, 1), float16],
%base_model.layer2.3.bn1.weight: Tensor[(128), float16],
%base_model.layer2.3.bn1.bias: Tensor[(128), float16],
%base_model.layer2.3.bn1.running_mean: Tensor[(128), float16],
%base_model.layer2.3.bn1.running_var: Tensor[(128), float16],
%base_model.layer2.3.conv2.weight: Tensor[(128, 128, 3, 3), float16],
%base_model.layer2.3.bn2.weight: Tensor[(128), float16],
%base_model.layer2.3.bn2.bias: Tensor[(128), float16],
%base_model.layer2.3.bn2.running_mean: Tensor[(128), float16],
%base_model.layer2.3.bn2.running_var: Tensor[(128), float16],
%base_model.layer2.3.conv3.weight: Tensor[(512, 128, 1, 1), float16],
%base_model.layer2.3.bn3.weight: Tensor[(512), float16],
%base_model.layer2.3.bn3.bias: Tensor[(512), float16],
%base_model.layer2.3.bn3.running_mean: Tensor[(512), float16],
%base_model.layer2.3.bn3.running_var: Tensor[(512), float16], %v796: Tensor[(1,
1, 64, 28, 28), float16], %v798: Tensor[(1, 1, 64, 28, 28), float16],
%base_model.layer3.0.conv1.net.weight: Tensor[(256, 512, 1, 1), float16],
%base_model.layer3.0.bn1.weight: Tensor[(256), float16],
%base_model.layer3.0.bn1.bias: Tensor[(256), float16],
%base_model.layer3.0.bn1.running_mean: Tensor[(256), float16],
%base_model.layer3.0.bn1.running_var: Tensor[(256), float16],
%base_model.layer3.0.conv2.weight: Tensor[(256, 256, 3, 3), float16],
%base_model.layer3.0.bn2.weight: Tensor[(256), float16],
%base_model.layer3.0.bn2.bias: Tensor[(256), float16],
%base_model.layer3.0.bn2.running_mean: Tensor[(256), float16],
%base_model.layer3.0.bn2.running_var: Tensor[(256), float16],
%base_model.layer3.0.conv3.weight: Tensor[(1024, 256, 1, 1), float16],
%base_model.layer3.0.bn3.weight: Tensor[(1024), float16],
%base_model.layer3.0.bn3.bias: Tensor[(1024), float16],
%base_model.layer3.0.bn3.running_mean: Tensor[(1024), float16],
%base_model.layer3.0.bn3.running_var: Tensor[(1024), float16],
%base_model.layer3.0.downsample.0.weight: Tensor[(1024, 512, 1, 1), float16],
%base_model.layer3.0.downsample.1.weight: Tensor[(1024), float16],
%base_model.layer3.0.downsample.1.bias: Tensor[(1024), float16],
%base_model.layer3.0.downsample.1.running_mean: Tensor[(1024), float16],
%base_model.layer3.0.downsample.1.running_var: Tensor[(1024), float16], %v862:
Tensor[(1, 1, 128, 14, 14), float16], %v864: Tensor[(1, 1, 128, 14, 14),
float16], %base_model.layer3.1.conv1.net.weight: Tensor[(256, 1024, 1, 1),
float16], %base_model.layer3.1.bn1.weight: Tensor[(256), float16],
%base_model.layer3.1.bn1.bias: Tensor[(256), float16],
%base_model.layer3.1.bn1.running_mean: Tensor[(256), float16],
%base_model.layer3.1.bn1.running_var: Tensor[(256), float16],
%base_model.layer3.1.conv2.weight: Tensor[(256, 256, 3, 3), float16],
%base_model.layer3.1.bn2.weight: Tensor[(256), float16],
%base_model.layer3.1.bn2.bias: Tensor[(256), float16],
%base_model.layer3.1.bn2.running_mean: Tensor[(256), float16],
%base_model.layer3.1.bn2.running_var: Tensor[(256), float16],
%base_model.layer3.1.conv3.weight: Tensor[(1024, 256, 1, 1), float16],
%base_model.layer3.1.bn3.weight: Tensor[(1024), float16],
%base_model.layer3.1.bn3.bias: Tensor[(1024), float16],
%base_model.layer3.1.bn3.running_mean: Tensor[(1024), float16],
%base_model.layer3.1.bn3.running_var: Tensor[(1024), float16], %v926:
Tensor[(1, 1, 128, 14, 14), float16], %v928: Tensor[(1, 1, 128, 14, 14),
float16], %base_model.layer3.2.conv1.net.weight: Tensor[(256, 1024, 1, 1),
float16], %base_model.layer3.2.bn1.weight: Tensor[(256), float16],
%base_model.layer3.2.bn1.bias: Tensor[(256), float16],
%base_model.layer3.2.bn1.running_mean: Tensor[(256), float16],
%base_model.layer3.2.bn1.running_var: Tensor[(256), float16],
%base_model.layer3.2.conv2.weight: Tensor[(256, 256, 3, 3), float16],
%base_model.layer3.2.bn2.weight: Tensor[(256), float16],
%base_model.layer3.2.bn2.bias: Tensor[(256), float16],
%base_model.layer3.2.bn2.running_mean: Tensor[(256), float16],
%base_model.layer3.2.bn2.running_var: Tensor[(256), float16],
%base_model.layer3.2.conv3.weight: Tensor[(1024, 256, 1, 1), float16],
%base_model.layer3.2.bn3.weight: Tensor[(1024), float16],
%base_model.layer3.2.bn3.bias: Tensor[(1024), float16],
%base_model.layer3.2.bn3.running_mean: Tensor[(1024), float16],
%base_model.layer3.2.bn3.running_var: Tensor[(1024), float16], %v990:
Tensor[(1, 1, 128, 14, 14), float16], %v992: Tensor[(1, 1, 128, 14, 14),
float16], %base_model.layer3.3.conv1.net.weight: Tensor[(256, 1024, 1, 1),
float16], %base_model.layer3.3.bn1.weight: Tensor[(256), float16],
%base_model.layer3.3.bn1.bias: Tensor[(256), float16],
%base_model.layer3.3.bn1.running_mean: Tensor[(256), float16],
%base_model.layer3.3.bn1.running_var: Tensor[(256), float16],
%base_model.layer3.3.conv2.weight: Tensor[(256, 256, 3, 3), float16],
%base_model.layer3.3.bn2.weight: Tensor[(256), float16],
%base_model.layer3.3.bn2.bias: Tensor[(256), float16],
%base_model.layer3.3.bn2.running_mean: Tensor[(256), float16],
%base_model.layer3.3.bn2.running_var: Tensor[(256), float16],
%base_model.layer3.3.conv3.weight: Tensor[(1024, 256, 1, 1), float16],
%base_model.layer3.3.bn3.weight: Tensor[(1024), float16],
%base_model.layer3.3.bn3.bias: Tensor[(1024), float16],
%base_model.layer3.3.bn3.running_mean: Tensor[(1024), float16],
%base_model.layer3.3.bn3.running_var: Tensor[(1024), float16], %v1054:
Tensor[(1, 1, 128, 14, 14), float16], %v1056: Tensor[(1, 1, 128, 14, 14),
float16], %base_model.layer3.4.conv1.net.weight: Tensor[(256, 1024, 1, 1),
float16], %base_model.layer3.4.bn1.weight: Tensor[(256), float16],
%base_model.layer3.4.bn1.bias: Tensor[(256), float16],
%base_model.layer3.4.bn1.running_mean: Tensor[(256), float16],
%base_model.layer3.4.bn1.running_var: Tensor[(256), float16],
%base_model.layer3.4.conv2.weight: Tensor[(256, 256, 3, 3), float16],
%base_model.layer3.4.bn2.weight: Tensor[(256), float16],
%base_model.layer3.4.bn2.bias: Tensor[(256), float16],
%base_model.layer3.4.bn2.running_mean: Tensor[(256), float16],
%base_model.layer3.4.bn2.running_var: Tensor[(256), float16],
%base_model.layer3.4.conv3.weight: Tensor[(1024, 256, 1, 1), float16],
%base_model.layer3.4.bn3.weight: Tensor[(1024), float16],
%base_model.layer3.4.bn3.bias: Tensor[(1024), float16],
%base_model.layer3.4.bn3.running_mean: Tensor[(1024), float16],
%base_model.layer3.4.bn3.running_var: Tensor[(1024), float16], %v1118:
Tensor[(1, 1, 128, 14, 14), float16], %v1120: Tensor[(1, 1, 128, 14, 14),
float16], %base_model.layer3.5.conv1.net.weight: Tensor[(256, 1024, 1, 1),
float16], %base_model.layer3.5.bn1.weight: Tensor[(256), float16],
%base_model.layer3.5.bn1.bias: Tensor[(256), float16],
%base_model.layer3.5.bn1.running_mean: Tensor[(256), float16],
%base_model.layer3.5.bn1.running_var: Tensor[(256), float16],
%base_model.layer3.5.conv2.weight: Tensor[(256, 256, 3, 3), float16],
%base_model.layer3.5.bn2.weight: Tensor[(256), float16],
%base_model.layer3.5.bn2.bias: Tensor[(256), float16],
%base_model.layer3.5.bn2.running_mean: Tensor[(256), float16],
%base_model.layer3.5.bn2.running_var: Tensor[(256), float16],
%base_model.layer3.5.conv3.weight: Tensor[(1024, 256, 1, 1), float16],
%base_model.layer3.5.bn3.weight: Tensor[(1024), float16],
%base_model.layer3.5.bn3.bias: Tensor[(1024), float16],
%base_model.layer3.5.bn3.running_mean: Tensor[(1024), float16],
%base_model.layer3.5.bn3.running_var: Tensor[(1024), float16], %v1182:
Tensor[(1, 1, 128, 14, 14), float16], %v1184: Tensor[(1, 1, 128, 14, 14),
float16], %base_model.layer4.0.conv1.net.weight: Tensor[(512, 1024, 1, 1),
float16], %base_model.layer4.0.bn1.weight: Tensor[(512), float16],
%base_model.layer4.0.bn1.bias: Tensor[(512), float16],
%base_model.layer4.0.bn1.running_mean: Tensor[(512), float16],
%base_model.layer4.0.bn1.running_var: Tensor[(512), float16],
%base_model.layer4.0.conv2.weight: Tensor[(512, 512, 3, 3), float16],
%base_model.layer4.0.bn2.weight: Tensor[(512), float16],
%base_model.layer4.0.bn2.bias: Tensor[(512), float16],
%base_model.layer4.0.bn2.running_mean: Tensor[(512), float16],
%base_model.layer4.0.bn2.running_var: Tensor[(512), float16],
%base_model.layer4.0.conv3.weight: Tensor[(2048, 512, 1, 1), float16],
%base_model.layer4.0.bn3.weight: Tensor[(2048), float16],
%base_model.layer4.0.bn3.bias: Tensor[(2048), float16],
%base_model.layer4.0.bn3.running_mean: Tensor[(2048), float16],
%base_model.layer4.0.bn3.running_var: Tensor[(2048), float16],
%base_model.layer4.0.downsample.0.weight: Tensor[(2048, 1024, 1, 1), float16],
%base_model.layer4.0.downsample.1.weight: Tensor[(2048), float16],
%base_model.layer4.0.downsample.1.bias: Tensor[(2048), float16],
%base_model.layer4.0.downsample.1.running_mean: Tensor[(2048), float16],
%base_model.layer4.0.downsample.1.running_var: Tensor[(2048), float16], %v1248:
Tensor[(1, 1, 256, 7, 7), float16], %v1250: Tensor[(1, 1, 256, 7, 7), float16],
%base_model.layer4.1.conv1.net.weight: Tensor[(512, 2048, 1, 1), float16],
%base_model.layer4.1.bn1.weight: Tensor[(512), float16],
%base_model.layer4.1.bn1.bias: Tensor[(512), float16],
%base_model.layer4.1.bn1.running_mean: Tensor[(512), float16],
%base_model.layer4.1.bn1.running_var: Tensor[(512), float16],
%base_model.layer4.1.conv2.weight: Tensor[(512, 512, 3, 3), float16],
%base_model.layer4.1.bn2.weight: Tensor[(512), float16],
%base_model.layer4.1.bn2.bias: Tensor[(512), float16],
%base_model.layer4.1.bn2.running_mean: Tensor[(512), float16],
%base_model.layer4.1.bn2.running_var: Tensor[(512), float16],
%base_model.layer4.1.conv3.weight: Tensor[(2048, 512, 1, 1), float16],
%base_model.layer4.1.bn3.weight: Tensor[(2048), float16],
%base_model.layer4.1.bn3.bias: Tensor[(2048), float16],
%base_model.layer4.1.bn3.running_mean: Tensor[(2048), float16],
%base_model.layer4.1.bn3.running_var: Tensor[(2048), float16], %v1312:
Tensor[(1, 1, 256, 7, 7), float16], %v1314: Tensor[(1, 1, 256, 7, 7), float16],
%base_model.layer4.2.conv1.net.weight: Tensor[(512, 2048, 1, 1), float16],
%base_model.layer4.2.bn1.weight: Tensor[(512), float16],
%base_model.layer4.2.bn1.bias: Tensor[(512), float16],
%base_model.layer4.2.bn1.running_mean: Tensor[(512), float16],
%base_model.layer4.2.bn1.running_var: Tensor[(512), float16],
%base_model.layer4.2.conv2.weight: Tensor[(512, 512, 3, 3), float16],
%base_model.layer4.2.bn2.weight: Tensor[(512), float16],
%base_model.layer4.2.bn2.bias: Tensor[(512), float16],
%base_model.layer4.2.bn2.running_mean: Tensor[(512), float16],
%base_model.layer4.2.bn2.running_var: Tensor[(512), float16],
%base_model.layer4.2.conv3.weight: Tensor[(2048, 512, 1, 1), float16],
%base_model.layer4.2.bn3.weight: Tensor[(2048), float16],
%base_model.layer4.2.bn3.bias: Tensor[(2048), float16],
%base_model.layer4.2.bn3.running_mean: Tensor[(2048), float16],
%base_model.layer4.2.bn3.running_var: Tensor[(2048), float16], %new_fc.weight:
Tensor[(60, 2048), float16], %new_fc.bias: Tensor[(60), float16]) {
%0 = reshape(%input, newshape=[8, 3, 224, 224]);
...
%499 = nn.avg_pool2d(%498, pool_size=[7, 7], padding=[0, 0, 0, 0]);
%500 = reshape(%499, newshape=[8, -1]);
%501 = nn.batch_flatten(%500);
%502 = multiply(1f, %501) an internal invariant was violated while
typechecking your program [16:23:08]
/home/data/git/tvm/src/relay/op/type_relations.cc:112: Check failed: t0->dtype
== t1->dtype (float32 vs. float16) :
; ;
%503 = nn.dense(%502, %new_fc.weight, units=60);
%504 = multiply(1f, %new_fc.bias) an internal invariant was violated while
typechecking your program [16:23:08]
/home/data/git/tvm/src/relay/op/type_relations.cc:112: Check failed: t0->dtype
== t1->dtype (float32 vs. float16) :
; ;
%505 = nn.bias_add(%503, %504);
%506 = reshape(%505, newshape=[1, 8, 60]);
%507 = mean(%506, axis=[1], keepdims=True);
squeeze(%507, axis=[1])
}
```
It seems the multiply op raised error. However, I do not know 1f means.
---
[Visit
Topic](https://discuss.tvm.ai/t/load-a-tvm-model-from-a-float16-onnx-model-error/6855/1)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https://discuss.tvm.ai/email/unsubscribe/6f4527e4b8ecda6c0dec1f6a2dd5bd0f71c74f22e3c8aa8d399a5ff0b95165c0).
