mxnet about safety-helmet-wearing-dataset HOT 26 OPEN

@njvisionpower 我关闭 CUDNN 就可以了。

from safety-helmet-wearing-dataset.

你好，麻烦问一下这个问题怎么解决啊
mxnet.base.MXNetError: [16:50:21] src/ndarray/ndarray.cc:1285: GPU is not enable

Download the correct version of mxnet by using "pip install mxnet-cu***" *** refers to your cuda version

from safety-helmet-wearing-dataset.

Thank you

from safety-helmet-wearing-dataset.

你好，再麻烦问一下这个问题怎么解决啊
OSError: libcudart.so.10.0: cannot open shared object file: No such file or directory

from safety-helmet-wearing-dataset.

麻烦问一下你用的是什么版本的mxnet、gluoncv?我安装的版本好像不太对，老是报错，哈哈

from safety-helmet-wearing-dataset.

Hi,

我用的mxnet1.5.0，gluoncv0.5.0。看你的报错显然是CUDA相关错误，比如CUDA是否安装正确、环境变量是否正确、所用CUDNN是否匹配等等，具体内容可以自己Google/Baidu下哈。

from safety-helmet-wearing-dataset.

import gluoncv
Traceback (most recent call last):
File "", line 1, in
File "build/bdist.linux-aarch64/egg/gluoncv/init.py", line 8, in
File "build/bdist.linux-aarch64/egg/gluoncv/data/init.py", line 7, in
File "build/bdist.linux-aarch64/egg/gluoncv/data/dataloader.py", line 9, in
ImportError: cannot import name _MultiWorkerIter

from safety-helmet-wearing-dataset.

@njvisionpower 请问您的 gluoncv 是什么版本，我这个出现这个问题

import gluoncv
Traceback (most recent call last):
File "", line 1, in
File "build/bdist.linux-aarch64/egg/gluoncv/init.py", line 8, in
File "build/bdist.linux-aarch64/egg/gluoncv/data/init.py", line 7, in
File "build/bdist.linux-aarch64/egg/gluoncv/data/dataloader.py", line 9, in
ImportError: cannot import name _MultiWorkerIter

from safety-helmet-wearing-dataset.

from safety-helmet-wearing-dataset.

from safety-helmet-wearing-dataset.

@yuanhanxiagn @liuwen-nj 有安装问题的话可以参考官方链接

from safety-helmet-wearing-dataset.

@njvisionpower 执行python test_symbol.py 时出现：
apache-mxnet-src-1.5.0-incubating/python/mxnet/gluon/block.py:1159: UserWarning: Cannot decide type for the following arguments. Consider providing them as input:
data: None
input_sym_arg_type = in_param.infer_type()[0]
[16:05:27] src/operator/nn/./cudnn/./cudnn_algoreg-inl.h:97: Running performance tests to find the best convolution algorithm, this can take a while... (set the environment variable MXNET_CUDNN_AUTOTUNE_DEFAULT to 0 to disable)

from safety-helmet-wearing-dataset.

这个打印无所谓，最终能执行输出结果就行，那个环境变量是挑选卷积的，不想用的话置0就可以了。

from safety-helmet-wearing-dataset.

@njvisionpower Traceback (most recent call last):
File "test_symbol.py", line 25, in
ax = utils.viz.cv_plot_bbox(img, bboxes[0], scores[0], box_ids[0], class_names=classes,thresh=0.4)
File "/usr/local/lib/python2.7/dist-packages/gluoncv/utils/viz/bbox.py", line 156, in cv_plot_bbox
bboxes = bboxes.asnumpy()
File "/home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/ndarray/ndarray.py", line 1996, in asnumpy
ctypes.c_size_t(data.size)))
File "/home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/base.py", line 253, in check_call
raise MXNetError(py_str(_LIB.MXGetLastError()))
mxnet.base.MXNetError: [14:33:01] src/operator/nn/./cudnn/cudnn_convolution-inl.h:683: Check failed: e == CUDNN_STATUS_SUCCESS (4 vs. 0) : cuDNN: CUDNN_STATUS_INTERNAL_ERROR
Stack trace:
[bt] (0) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x4c) [0x7f72e72a44]
[bt] (1) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::op::CuDNNConvolutionOp::CuDNNAlgoSetter(mxnet::RunContext const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, cudnnDataType_t, cudnnDataType_t, mxnet::op::CuDNNAlgo<cudnnConvolutionFwdAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdDataAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdFilterAlgo_t>)+0x350) [0x7f76aa43c8]
[bt] (2) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(std::_Function_handler<void (mxnet::op::CuDNNAlgo<cudnnConvolutionFwdAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdDataAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdFilterAlgo_t>), mxnet::op::CuDNNConvolutionOp::SelectAlgo(mxnet::RunContext const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, cudnnDataType_t, cudnnDataType_t)::{lambda(mxnet::op::CuDNNAlgo<cudnnConvolutionFwdAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdDataAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdFilterAlgo_t>)#1}>::_M_invoke(std::_Any_data const&, mxnet::op::CuDNNAlgo<cudnnConvolutionFwdAlgo_t>&&, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdDataAlgo_t>&&, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdFilterAlgo_t>&&)+0x40c) [0x7f76aa9e54]
[bt] (3) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::op::CuDNNAlgoRegmxnet::op::ConvolutionParam::FindOrElseRegister(mxnet::op::ConvolutionParam const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, cudnnDataType_t, cudnnDataType_t, cudnnDataType_t, int, bool, mxnet::op::CuDNNAlgo<cudnnConvolutionFwdAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdDataAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdFilterAlgo_t>, std::function<void (mxnet::op::CuDNNAlgo<cudnnConvolutionFwdAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdDataAlgo_t>, mxnet::op::CuDNNAlgo<cudnnConvolutionBwdFilterAlgo_t>)> const&)+0x32c) [0x7f76aac1b4]
[bt] (4) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::op::CuDNNConvolutionOp::SelectAlgo(mxnet::RunContext const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, std::vector<mxnet::TShape, std::allocatormxnet::TShape > const&, cudnnDataType_t, cudnnDataType_t)+0x160) [0x7f76aac4d8]
[bt] (5) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(+0x4d436bc) [0x7f76a856bc]
[bt] (6) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(void mxnet::op::ConvolutionComputemshadow::gpu(nnvm::NodeAttrs const&, mxnet::OpContext const&, std::vector<mxnet::TBlob, std::allocatormxnet::TBlob > const&, std::vector<mxnet::OpReqType, std::allocatormxnet::OpReqType > const&, std::vector<mxnet::TBlob, std::allocatormxnet::TBlob > const&)+0x4f0) [0x7f76a89000]
[bt] (7) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::imperative::PushFCompute(std::function<void (nnvm::NodeAttrs const&, mxnet::OpContext const&, std::vector<mxnet::TBlob, std::allocatormxnet::TBlob > const&, std::vector<mxnet::OpReqType, std::allocatormxnet::OpReqType > const&, std::vector<mxnet::TBlob, std::allocatormxnet::TBlob > const&)> const&, nnvm::Op const*, nnvm::NodeAttrs const&, mxnet::Context const&, std::vector<mxnet::engine::Var*, std::allocatormxnet::engine::Var* > const&, std::vector<mxnet::engine::Var*, std::allocatormxnet::engine::Var* > const&, std::vector<mxnet::Resource, std::allocatormxnet::Resource > const&, std::vector<mxnet::NDArray*, std::allocatormxnet::NDArray* > const&, std::vector<mxnet::NDArray*, std::allocatormxnet::NDArray* > const&, std::vector<unsigned int, std::allocator > const&, std::vector<mxnet::OpReqType, std::allocatormxnet::OpReqType > const&)::{lambda(mxnet::RunContext)#1}::operator()(mxnet::RunContext) const+0x194) [0x7f74d7dfec]
[bt] (8) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(std::_Function_handler<void (mxnet::RunContext), mxnet::imperative::PushFCompute(std::function<void (nnvm::NodeAttrs const&, mxnet::OpContext const&, std::vector<mxnet::TBlob, std::allocatormxnet::TBlob > const&, std::vector<mxnet::OpReqType, std::allocatormxnet::OpReqType > const&, std::vector<mxnet::TBlob, std::allocatormxnet::TBlob > const&)> const&, nnvm::Op const*, nnvm::NodeAttrs const&, mxnet::Context const&, std::vector<mxnet::engine::Var*, std::allocatormxnet::engine::Var* > const&, std::vector<mxnet::engine::Var*, std::allocatormxnet::engine::Var* > const&, std::vector<mxnet::Resource, std::allocatormxnet::Resource > const&, std::vector<mxnet::NDArray*, std::allocatormxnet::NDArray* > const&, std::vector<mxnet::NDArray*, std::allocatormxnet::NDArray* > const&, std::vector<unsigned int, std::allocator > const&, std::vector<mxnet::OpReqType, std::allocatormxnet::OpReqType > const&)::{lambda(mxnet::RunContext)#1}>::_M_invoke(std::_Any_data const&, mxnet::RunContext&&)+0x28) [0x7f74d7e1a8]

[14:33:01] src/resource.cc:279: Ignore CUDA Error [14:33:01] src/storage/./pooled_storage_manager.h:97: CUDA: unknown error
Stack trace:
[bt] (0) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x4c) [0x7f72e72a44]
[bt] (1) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::storage::GPUPooledStorageManager::DirectFreeNoLock(mxnet::Storage::Handle)+0xdc) [0x7f7540bb14]
[bt] (2) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::storage::GPUPooledStorageManager::DirectFree(mxnet::Storage::Handle)+0x54) [0x7f7540e46c]
[bt] (3) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::StorageImpl::DirectFree(mxnet::Storage::Handle)+0x94) [0x7f75406524]
[bt] (4) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(std::_Function_handler<void (mxnet::RunContext), mxnet::resource::ResourceManagerImpl::ResourceTempSpace<(mxnet::ResourceRequest::Type)1>::~ResourceTempSpace()::{lambda(mxnet::RunContext)#1}>::_M_invoke(std::_Any_data const&, mxnet::RunContext&&)+0x88) [0x7f7551e018]
[bt] (5) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(+0x36ac10c) [0x7f753ee10c]
[bt] (6) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::engine::ThreadedEngine::ExecuteOprBlock(mxnet::RunContext, mxnet::engine::OprBlock*)+0x820) [0x7f753e8bc0]
[bt] (7) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::engine::ThreadedEnginePerDevice::PushToExecute(mxnet::engine::OprBlock*, bool)+0xa0) [0x7f754021e8]
[bt] (8) /home/ubuntu/Safehat/mxnet-tar/apache-mxnet-src-1.5.0-incubating/python/mxnet/../../lib/libmxnet.so(mxnet::engine::ThreadedEngine::Push(mxnet::engine::Opr*, mxnet::Context, int, bool)+0x360) [0x7f753f1178]

from safety-helmet-wearing-dataset.

@njvisionpower 提示这个错误，是不是安装环境的问题？

from safety-helmet-wearing-dataset.

将老哥的数据集放在ssd-tensorflow 上跑，效果极差，不知道 @njvisionpower 有什么想法没，运行1w step了，map只有0.25

from safety-helmet-wearing-dataset.

@Single430 你跑的1w个step batch size用的多大，还有就是输入大小和网络用的什么？我用gluoncv的SSD resnet 512也train过，map虽然没有YOLOv3高，但也低不了几个点。

from safety-helmet-wearing-dataset.

BS 32， 输入300x300，vgg16预训练模型 @njvisionpower

from safety-helmet-wearing-dataset.

使用了您的模型，用测试算了mAP 效果没有达到0.88，用的是facebook提供的voc_eval.py来计算的, 是有什么其它因素吗？ @njvisionpower
network = "yolo3_darknet53_voc"
hat: P 0.97 R 0.4 AP 0.76
person: P 0.97 R 0.4 AP 0.76

mAP : 0.76 < 0.88

from safety-helmet-wearing-dataset.

@Single430 R太低了，我用gluoncv自带的map计算方法，训练的时候二者至少八十几

from safety-helmet-wearing-dataset.

@njvisionpower 请问有没有caffe模型？谢谢

from safety-helmet-wearing-dataset.

@Single430 看训练代码作者用的是voc07metric，voc_eval.py应该用的是voc10的标准，两个AP计算方式有不同，但是R这么低应该是有问题的

from safety-helmet-wearing-dataset.

我这个是在test.txt文件上测的，voc_eval.py是官方提供的，而且里面可以选择用voc07标准，不过可能是我什么地方不对吧，请问你是怎么算的？ @VectXmy

from safety-helmet-wearing-dataset.

你好，麻烦问一下这个问题怎么解决啊
mxnet.base.MXNetError: [16:50:21] src/ndarray/ndarray.cc:1285: GPU is not enable

您好，我也是和您一样的问题，我把版本改成0.5和1.5的还是报相同的错误，请问您解决了吗

from safety-helmet-wearing-dataset.

hello,下面这个报错怎么解决呢（运行train_yolo.py）
Traceback (most recent call last):
File "E:/pycharm/.idea/opencv/saftyhelmet/train_yolo.py", line 335, in
net = get_model(net_name, pretrained_base=True, classes=classes)
File "E:\pycharm.idea\opencv\venv\lib\site-packages\gluoncv\model_zoo\model_zoo.py", line 361, in get_model
net = _modelsname
File "E:\pycharm.idea\opencv\venv\lib\site-packages\gluoncv\model_zoo\yolo\yolo3.py", line 601, in yolo3_darknet53_voc
pretrained=pretrained_base, norm_layer=norm_layer, norm_kwargs=norm_kwargs, **kwargs)
File "E:\pycharm.idea\opencv\venv\lib\site-packages\gluoncv\model_zoo\yolo\darknet.py", line 187, in darknet53
return get_darknet('v3', 53, **kwargs)
File "E:\pycharm.idea\opencv\venv\lib\site-packages\gluoncv\model_zoo\yolo\darknet.py", line 165, in get_darknet
'darknet%d'%(num_layers), tag=pretrained, root=root), ctx=ctx)
File "E:\pycharm.idea\opencv\venv\lib\site-packages\mxnet\gluon\block.py", line 555, in load_parameters
params[name]._load_init(loaded[name], ctx, cast_dtype=cast_dtype, dtype_source=dtype_source)
File "E:\pycharm.idea\opencv\venv\lib\site-packages\mxnet\gluon\parameter.py", line 282, in _load_init
self.name, str(self.shape), str(data.shape))
AssertionError: Failed loading Parameter 'darknetv30_dense0_bias' from saved params: shape incompatible expected (['hat', 'person'],) vs saved (1000,)

from safety-helmet-wearing-dataset.

Hello ,
tel me please how can be change the color of detection ,
i doesn't find the line must be edited to chose the color ;

like you see in images , the person who wearing hat must be draw with green color and the person who doesn't wearing hat must be draw with red color

from safety-helmet-wearing-dataset.