Tag Archives: Deep learning

[Solved] error: (-2:Unspecified error) The function is not implemented. Rebuild the library with Windows

Error: error: (-2:Unspecified error) The function is not implemented. Rebuild the library with Windows

I’m on windows, and when running an image motion blur algorithm from the cv-python library, this error is suddenly reported.
Searching for similar problems has said that the path to read the image is wrong, and reinstalling a higher version of the cv library, but not all of them are useless.
All the functions under the cv2 class are not working.

Solution:
First:

pip uninstall opencv-python

Then:

pip install opencv-python

Just reinstall this package. I don’t know the cause. The problem occurred after I installed the evaluations package. It should be caused by the conflict between versions
stackoverflow: https://stackoverflow.com/questions/67120450/error-2unspecified-error-the-function-is-not-implemented-rebuild-the-libra

[Solved] TVM operate error: TVMError: AssertionError

TVM operate error: TVMError: AssertionError

Traceback (most recent call last):
  File "tune_relay_x86.py", line 248, in <module>
    tune_and_evaluate(tuning_option)
  File "tune_relay_x86.py", line 241, in tune_and_evaluate
    lib = relay.build_module.build(mod, target=target, params=params)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/relay/build_module.py", line 468, in build
    graph_json, runtime_mod, params = bld_mod.build(
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/relay/build_module.py", line 196, in build
    self._build(mod, target, target_host, executor, runtime, workspace_memory_pools, mod_name)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/_ffi/_ctypes/packed_func.py", line 237, in __call__
    raise get_last_ffi_error()
tvm._ffi.base.TVMError: Traceback (most recent call last):
  30: TVMFuncCall
  29: 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
  28: tvm::relay::backend::RelayBuildModule::BuildRelay(tvm::IRModule, tvm::runtime::String const&)
  27: tvm::relay::backend::RelayBuildModule::OptimizeImpl(tvm::IRModule)
  26: tvm::transform::Pass::operator()(tvm::IRModule) const
  25: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
  24: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
  23: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
  22: tvm::relay::transform::FunctionPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
  21: _ZN3tvm7runtime13PackedFuncObj9ExtractorINS0_16PackedFuncSubObjIZNS0_15TypedPackedFuncIFNS_5relay8FunctionES6_NS_8IRModuleENS_9transform11PassContextEEE17AssignTypedLambdaIZNS5_9transform13AlterOpLayoutEvEUlS6_S7_S9_E_EEvT_EUlRKNS0_7TVMArgsEPNS0_11TVMRetValueEE_EEE4CallEPKS1_SG_SK_
  20: tvm::relay::alter_op_layout::AlterOpLayout(tvm::RelayExpr const&)
  19: tvm::relay::ForwardRewrite(tvm::RelayExpr const&, tvm::runtime::TypedPackedFunc<tvm::RelayExpr (tvm::relay::Call const&, tvm::runtime::Array<tvm::RelayExpr, void> const&, tvm::runtime::ObjectRef const&)> const&, std::function<tvm::runtime::ObjectRef (tvm::relay::Call const&)>, std::function<tvm::RelayExpr (tvm::RelayExpr const&)>)
  18: tvm::relay::MixedModeMutator::VisitExpr(tvm::RelayExpr const&)
  17: tvm::relay::MixedModeMutator::VisitLeaf(tvm::RelayExpr const&)
  16: _ZN3tvm5relay16MixedModeMutator17DispatchVisitExprERKNS_9Re
  15: tvm::relay::ExprMutator::VisitExpr(tvm::RelayExpr const&)
  14: tvm::relay::ExprFunctor<tvm::RelayExpr (tvm::RelayExpr const&)>::VisitExpr(tvm::RelayExpr const&)
  13: _ZZN3tvm5relay11ExprFunctorIFNS_9RelayExprERKS2_EE10InitVTableEvENUlR
  12: tvm::relay::ExprMutator::VisitExpr_(tvm::relay::FunctionNode const*)
  11: tvm::relay::MixedModeMutator::VisitExpr(tvm::RelayExpr const&)
  10: tvm::relay::MixedModeMutator::VisitLeaf(tvm::RelayExpr const&)
  9: _ZN3tvm5relay16MixedModeMutator17DispatchVisitExprERKNS_9Re
  8: tvm::relay::ExprMutator::VisitExpr(tvm::RelayExpr const&)
  7: tvm::relay::ExprFunctor<tvm::RelayExpr (tvm::RelayExpr const&)>::VisitExpr(tvm::RelayExpr const&)
  6: _ZZN3tvm5relay11ExprFunctorIFNS_9RelayExprERKS2_EE10InitVTableEvENUlR
  5: tvm::relay::MixedModeMutator::VisitExpr_(tvm::relay::CallNode const*)
  4: tvm::relay::ForwardRewriter::Rewrite_(tvm::relay::CallNode const*, tvm::RelayExpr const&)
  3: tvm::runtime::PackedFuncObj::Extractor<tvm::runtime::PackedFuncSubObj<tvm::runtime::TypedPackedFunc<tvm::RelayExpr (tvm::relay::Call const&, tvm::runtime::Array<tvm::RelayExpr, void> const&, tvm::runtime::ObjectRef const&)>::AssignTypedLambda<tvm::RelayExpr (*)(tvm::relay::Call const&, tvm::runtime::Array<tvm::RelayExpr, void> const&, tvm::runtime::ObjectRef const&)>(tvm::RelayExpr (*)(tvm::relay::Call const&, tvm::runtime::Array<tvm::RelayExpr, void> const&, tvm::runtime::ObjectRef const&))::{lambda(tvm::runtime::TVMArgs const&, tvm::runtime::TVMRetValue*)#1}> >::Call(tvm::runtime::PackedFuncObj const*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)
  2: tvm::RelayExpr tvm::relay::LayoutRewriter<tvm::relay::alter_op_layout::AlterTransformMemorizer>(tvm::relay::Call const&, tvm::runtime::Array<tvm::RelayExpr, void> const&, tvm::runtime::ObjectRef const&)
  1: tvm::relay::alter_op_layout::AlterTransformMemorizerNode::CallWithNewLayouts(tvm::relay::Call const&, tvm::Attrs, std::vector<tvm::RelayExpr, std::allocator<tvm::RelayExpr> > const&)
  0: tvm::runtime::PackedFuncObj::Extractor<tvm::runtime::PackedFuncSubObj<TVMFuncCreateFromCFunc::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#2}> >::Call(tvm::runtime::PackedFuncObj const*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/_ffi/_ctypes/packed_func.py", line 81, in cfun
    rv = local_pyfunc(*pyargs)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/relay/op/nn/_nn.py", line 112, in alter_op_layout_dense
    return topi.nn.dense_alter_layout(attrs, inputs, tinfos, out_type)
  File "/home/lizhenxu/anaconda3/lib/python3.8/site-packages/decorator.py", line 231, in fun
    return caller(func, *(extras + args), **kw)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/target/generic_func.py", line 286, in dispatch_func
    return dispatch_dict[k](*args, **kwargs)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/topi/x86/dense_alter_op.py", line 51, in _alter_dense_layout
    _, outs = relay.backend.te_compiler.select_implementation(
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/relay/backend/te_compiler.py", line 201, in select_implementation
    outs = impl.compute(attrs, inputs, out_type)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/relay/op/op.py", line 126, in compute
    return _OpImplementationCompute(self, attrs, inputs, out_type)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/_ffi/_ctypes/packed_func.py", line 237, in __call__
    raise get_last_ffi_error()
  3: TVMFuncCall
  2: tvm::runtime::PackedFuncObj::Extractor<tvm::runtime::PackedFuncSubObj<tvm::relay::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#4}> >::Call(tvm::runtime::PackedFuncObj const*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)
  1: tvm::relay::OpImplementation::Compute(tvm::Attrs const&, tvm::runtime::Array<tvm::te::Tensor, void> const&, tvm::Type const&)
  0: tvm::runtime::PackedFuncObj::Extractor<tvm::runtime::PackedFuncSubObj<TVMFuncCreateFromCFunc::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#2}> >::Call(tvm::runtime::PackedFuncObj const*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/_ffi/_ctypes/packed_func.py", line 81, in cfun
    rv = local_pyfunc(*pyargs)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/relay/op/strategy/generic.py", line 833, in _compute_dense
    return [topi_compute(*args)]
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/autotvm/task/topi_integration.py", line 164, in wrapper
    cfg = DispatchContext.current.query(tgt, workload)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/autotvm/task/dispatcher.py", line 76, in query
    ret = self._query_inside(target, workload)
  File "/home/lizhenxu/ZJJ/tvm_0.9/tvm/python/tvm/autotvm/task/dispatcher.py", line 421, in _query_inside
    assert wkl == workload
TVMError: AssertionError

Run the official website code tune_relay_x86.py report the error above.

Solution: delete the .log file that was generated by running this code before

Pytorch Error: error: identifier “AT_CHECK“ is undefined [How to Solve]

These two days, run a mask RCNN series of network code, and execute the command Python 3 setup A bunch of such errors are received during py build development, as follows:

.../detectron2/layers/csrc/deformable/deform_conv.h(136): error: identifier "AT_CHECK" is undefined

.../detectron2/layers/csrc/deformable/deform_conv.h(184): error: identifier "AT_CHECK" is undefined

.../detectron2/layers/csrc/deformable/deform_conv.h(234): error: identifier "AT_CHECK" is undefined

.../detectron2/layers/csrc/deformable/deform_conv.h(284): error: identifier "AT_CHECK" is undefined

.../detectron2/layers/csrc/deformable/deform_conv.h(341): error: identifier "AT_CHECK" is undefined

.../detectron2/layers/csrc/deformable/deform_conv_cuda.cu(155): error: identifier "AT_CHECK " is undefined

.../detectron2/layers/csrc/deformable/deform_conv_cuda.cu(338): error: identifier "AT_CHECK " is undefined

.../detectron2/layers/csrc/deformable/deform_conv_cuda.cu(503): error: identifier "AT_CHECK " is undefined

.../detectron2/layers/csrc/deformable/deform_conv_cuda.cu(696): error: identifier "AT_CHECK " is undefined

.../detectron2/layers/csrc/deformable/deform_conv_cuda.cu(823): error: identifier "AT_CHECK " is undefined

.../detectron2/layers/csrc/deformable/deform_conv_cuda.cu(953): error: identifier "AT_CHECK " is undefined

11 errors detected in the compilation of ".../detectron2/layers/csrc/deformable/deform_conv _cuda.cu".

Solution:

Replace all AT_CHECK in deform_conv_cuda.cpp with TORCH_CHECK

or macro definition:

#ifndef AT_CHECK
#define AT_CHECK TORCH_CHECK 
#endif

[DL Common Issue] RuntimeError: CUDA error 59: Device-side assert triggered

Problem: Run into a CUDA error during training
Solutions:
This error occurs due to the following two reasons:

    Inconsistency between the number of labels/classes and the number of output unitsThe input of the loss function may be incorrect

In my case the error occurs as the loss function is not correctly chosen : change to nn.BCELoss() from nn.CrossEntropyLoss()
Reference: https://towardsdatascience.com/cuda-error-device-side-assert-triggered-c6ae1c8fa4c3

[Solved] Model training Error: _pickle.PicklingError: Can’t pickle

How to Solve Model training Error: _pickle.Picklingerror: can’t pickle

 

1. Problem description

Recently, when learning the target tracking model of siamfc model, it is found that the following problems always occur during model training on window platform:

_pickle.PicklingError: Can’t pickle <class ‘pairwise.GenericDict’>: attribute lookup GenericDict on pairwise failed

See the following figure for details:

2. Solution

The main problem is that the code is written on Linux platform, test.py has no problem in actual operation, but train.py has problem in window platform, the main problem is on Dataloader, so we can modify this part of the code. The main problem lies in the Dataloader, so we can modify this part of the code. Or we can directly train and test the model under Linux with the source code.

The solution to this problem on Window platform is as follows:

Modify the original code from num_workers = 4 to num_workers = 0 and it will work as follows.

after the modification is completed, the operation effect is as follows:

[Solved] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter st

When using tensorflow.keras, this error is often reported during model training:

tensorflow/core/kernels/data/generator_dataset_op.cc:107] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated.
	 [[{{node PyFunc}}]]

tensorflow/core/kernels/data/generator_dataset_op.cc:107] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated.

       [[{{node PyFunc}}]]

 

According to my own experience, there are several reasons for this errory:

1. The input image_size and input_shape does not match or is not defined when the model is built. Note that the input_shape must be defined when defining the first layer of the convolutional layer, e.g.

    model = keras.models.Sequential([
        # Input image [None,224,224,3]
        # Convolution layer 1: 32 5*5*3 filters, step size set to 1, fill set to same
        # Output [None,32,32,3]
        keras.layers.Conv2D(32, kernel_size=5, strides=1, padding='same', data_format='channels_last',
                            activation='relu', input_shape=(224, 224, 3)),

2. There is also train_generator and validate_generator related parameters must be consistent, such as batch_size, target_size, class_mode, etc.

3. The configuration limit itself, try to change the batch_size to a smaller size, or even to 1

4. The last program did not finish completely, finish all python programs to see.

[Solved] AttributeError: ‘_IncompatibleKeys‘ object has no attribute ‘parameters‘

Errors are reported when running the pytorch program. It should be a problem with the torch syntax.

Original code:

model=CNN（）#CNN as a self-compiling neural network model
best_model_wts = copy.deepcopy(model.state_dict())
*************************************#（the error codes is below）
model=model.load_state_dict(best_model_wts)

Error message:

Modified code:

model=CNN（）#CNN as a self-compiling neural network model
best_model_wts = copy.deepcopy(model.state_dict())
model.load_state_dict(best_model_wts)

[Solved] mmdetection Error: ImportError: /home/user/repos/mmdetection/mmdet/ops/dcn/deform_conv_cuda.cpython-37m-x

Environment configuration: torch 1.11.0 + CUDA 11.3 (latest)

Use mmdetection to infer:

from mmdet.apis import init_detector, inference_detector

Errors are reported as follows:

ImportError: /home/user/repos/mmdetection/mmdet/ops/dcn/deform_conv_cuda.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN6caffe26detail37_typeMetaDataInstance_preallocated_32E

 

The problem has been solved. The reason for the error is that the pytorch version is too new. Although openmmlab supports the latest version, it will still cause the error.

Solution:

Degrade the pytorch version to torch 1.6.0 + cu102, query the official GitHub of openmmlab, uninstall and reinstall mmcv 1.3.9, and re run the mmdetection code to solve the error.

[Solved] kitt2bag Error: Failed to find match for field intensity

Problem analysis: this problem usually occurs when the point cloud type with intensity field (such as PointXYZI) is used to load the point cloud without intensity information. The point cloud data structure downloaded from Kitti data set is (x, y, Z, I) which contains intensity information, so the above problem may be caused by the loss of point cloud intensity information when bin file is converted to bag file?

Solution:

  1. Modify kitti2bag.
  2. Generate new bag.

Modification steps:

enter whereis kitti2bag in the terminal to find the path of the file.

whereis kitti2bag

Open the file vim/gedit in this path and change ‘i’ to ‘intensity’

Modified
        # fill pcl msg
        fields = [PointField('x', 0, PointField.FLOAT32, 1),
                  PointField('y', 4, PointField.FLOAT32, 1),
                  PointField('z', 8, PointField.FLOAT32, 1),
                  PointField('intensity', 12, PointField.FLOAT32, 1)]
        pcl_msg = pcl2.create_cloud(header, fields, scan)

#Before
        # fill pcl msg
        fields = [PointField('x', 0, PointField.FLOAT32, 1),
                  PointField('y', 4, PointField.FLOAT32, 1),
                  PointField('z', 8, PointField.FLOAT32, 1),
                  PointField('i', 12, PointField.FLOAT32, 1)]
        pcl_msg = pcl2.create_cloud(header, fields, scan)i

Re-convert bag file.

[Solved] Torch Build Module Error: NotImplementedError

It’s probably such an error reporting method. I’ve been using torch for so many years. I first encountered this error NotImplementedError
I’m not using a nightly version

Traceback (most recent call last):

  File "xxxxx\x.py", line 268, in <module>
    print(x(y).shape)

  File "xxxxx\lib\site-packages\torch\nn\modules\module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)

  File "xxxxx\x.py", line 259, in forward
    x = self.features(x)

  File "xxxxx\lib\site-packages\torch\nn\modules\module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)

  File "xxxxx\lib\site-packages\torch\nn\modules\container.py", line 119, in forward
    input = module(input)

  File "xxxxx\lib\site-packages\torch\nn\modules\module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)

  File "xxxxx\lib\site-packages\torch\nn\modules\module.py", line 201, in _forward_unimplemented
    raise NotImplementedError

NotImplementedError

Call self.forward in _call_impl

result = self.forward(*input, **kwargs)

If you inherit nn.Module, and if you don’t implement self.forward, it will

raise NotImplementedError

It turns out that when I use this function, I really don’t have the forward method:

class Hswish(nn.Module):

    def __init__(self, inplace=True):
        super(Hswish, self).__init__()
        self.inplace = inplace

    def __swish(self, x, beta, inplace=True):
        # But this swish is not used by H-swish
        # The reason it's called H-swish is to make the sigmoid hard
        # approximated by Relu6(x+3)/6
        # Reduced computational effort for embedded deployment
        return x * F.sigmoid(beta * x, inplace)

    @staticmethod
    def Hsigmoid(x, inplace=True):
        return F.relu6(x + 3, inplace=inplace)/6

    def foward(self, x):
        return x * self.Hsigmoid(x, self.inplace)

forward Write as foward …

torch.max Example (How to Use)

torch.max(input, dim)

pred = torch.max(input, dim)

Returns the maximum value per row (dim = 1) or column (dim = 0).

_, pred = torch.max(input, dim)

Only the position of the maximum value in each row (dim = 1) or column (dim = 0) is returned.

Example:

import torch

# Construct a 5x3 randomly initialized matrix
x = torch.rand(5, 3)
print('input: ', x)
print('-'*10)
y1 = torch.max(x, 1)
print('max by row: ', y1)
print('-'*10)
y2 = torch.max(x, 0)
print('max by col: ', y2)
print('-'*10)
_, y3 = torch.max(x, 1)
print('max index by row: ', y3)
print('-'*10)
_, y4 = torch.max(x, 0)
print('max index by col: ', y4)

Output result:

input:  tensor([[0.5504, 0.3160, 0.2448],
        [0.8694, 0.3295, 0.2085],
        [0.5530, 0.9984, 0.3531],
        [0.2874, 0.1025, 0.9419],
        [0.0867, 0.4234, 0.8334]])
----------
max by row:  torch.return_types.max(
values=tensor([0.5504, 0.8694, 0.9984, 0.9419, 0.8334]),
indices=tensor([0, 0, 1, 2, 2]))
----------
max by col:  torch.return_types.max(
values=tensor([0.8694, 0.9984, 0.9419]),
indices=tensor([1, 2, 3]))
----------
max index by row:  tensor([0, 0, 1, 2, 2])
----------
max index by col:  tensor([1, 2, 3])