gjy3035 / pcc-net Goto Github PK

你好!
因为我在训练到600多epoch的时候报错了,但是用“tensorboard”也能输出所有的mae, mse包括test数据集的mae, mse.
这里我只执行了python train_lr.py, 没有执行python test.py, 所以只执行python train_lr.py输出的mae, mse具体指的什么mae和mse, 我是必须要执行test才是输出的test的mae和mse?

在test怎么样才能预测一张具体的图片的人数呢

你好,
__C.DATA.DATA_PATH = '/media/D/DataSet/CC/' +str(__C.DATA.STD_SIZE[0]) + 'x' + str(__C.DATA.STD_SIZE[1]) + '/shanghaitech_part_B'
这里的/media/D/DataSet/CC/对应本地的什么路径? 比如这个项目叫cnn, cnn下有个文件夹dataset,这里怎么对应

Please, in data preparation, how did you resize the locations of key points???

你好!
因为我在训练到600多epoch的时候报错了,但是用tensorboard也能输出所有的mae, mse包括test数据集的mae, mse,但是我只执行了python train_lr.py, 没有执行python test.py, 所以只执行python train_lr.py输出的mae, mse具体指的什么mae和mse, 我是必须要执行test才是输出的test的mae和mse?

hello author,when I am training the shanghai_dataset model ,I meet this problem,what is maybe the reason of it? The error likes this:
Traceback (most recent call last):
File "train_lr.py", line 232, in
main()
File "train_lr.py", line 67, in main
i_tb,model_path = train(train_loader, net, optimizer, epoch, i_tb)
File "train_lr.py", line 96, in train
pred_map,pred_cls, pred_seg = net(img, gt_map, roi, gt_roi, gt_seg)
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torch/nn/modules/module.py", line 541, in call
result = self.forward(*input, **kwargs)
File "/home/zhulei/PCC-Net/models/CC.py", line 34, in forward
density_map, density_cls_score,pred_seg = self.CCN(img,roi)
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torch/nn/modules/module.py", line 541, in call
result = self.forward(*input, **kwargs)
File "/home/zhulei/PCC-Net/models/ori_big.py", line 80, in forward
x_bbx = self.roi_pool(x_hlp, roi)
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torch/nn/modules/module.py", line 541, in call
result = self.forward(*input, **kwargs)
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torchvision/ops/roi_pool.py", line 75, in forward
return roi_pool(input, rois, self.output_size, self.spatial_scale)
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torchvision/ops/roi_pool.py", line 62, in roi_pool
return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torchvision/ops/roi_pool.py", line 19, in forward
_C = _lazy_import()
File "/home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torchvision/extension.py", line 12, in _lazy_import
from torchvision import _C as C
ImportError: /home/zhulei/anaconda3/envs/maskrcnn_benchmark/lib/python3.6/site-packages/torchvision/_C.so: undefined symbol: _ZN3c107Warning4warnENS_14SourceLocationENSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE

是用验证集验证的时间吗？ 还是有其他的计算方法？
另外还想问，如果原来的model 输出的结果是原图的1/4大小，该怎么计算时间呢？

thank you for your code. may i get this paper? or it's not publish
and can you provide the code of data preparation or some information about how to get seg data

Thanks for your work.
How can I use your code to count people on a single image ?

你好
min_gt_count = 2.7126654189217972e-05
max_gt_count = 0.001306603490202515

wts = torch.FloatTensor(
[ 0.10194444, 0.07416667, 0.08361111, 0.09277778, 0.10388889,
0.10416667, 0.10805556, 0.11 , 0.11111111, 0.11027778]
)
__C.TRAIN.INPUT_SIZE = (512,680)
这些初始值是怎么得来的呢?

I am trying to run this model on another dataset, but I am confused about how to generate the segment map according to the density map with the csv file? Thank you

I want to generate the value of each class and have an output for the classification task, any idea?!

你好!
我训练模型的时候,训练到634epoch的时候,遇到报错,
Traceback (most recent call last):
File "train_lr.py", line 234, in
main()
File "train_lr.py", line 69, in main
i_tb,model_path = train(train_loader, net, optimizer, epoch, i_tb)
File "train_lr.py", line 82, in train
for i, data in enumerate(train_loader, 0):
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/dataloader.py", line 345, in next
data = self._next_data()
这里有可能是什么原因造成的呢

你好
因为我在训练到600多epoch的时候报错了,但是用tensorboard也能输出所有的mae, mse包括test数据集的mae, mse,但是我只执行了python train_lr.py, 没有执行python test.py, 所以只执行python train_lr.py输出的mae, mse具体指的什么mae和mse, 我是必须要执行test才是输出的test的mae和mse?

Hi man, thanks for your share!

An error occurred when I run python build.py, the error is as follows:

Traceback (most recent call last):
  File "build.py", line 30, in <module>
    extra_objects=extra_objects
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/torch/utils/ffi/__init__.py", line 159, in create_extension
    ffi.cdef(_typedefs + all_headers_source)
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/cffi/api.py", line 112, in cdef
    self._cdef(csource, override=override, packed=packed, pack=pack)
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/cffi/api.py", line 126, in _cdef
    self._parser.parse(csource, override=override, **options)
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/cffi/cparser.py", line 347, in parse
    self._internal_parse(csource)
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/cffi/cparser.py", line 352, in _internal_parse
    ast, macros, csource = self._parse(csource)
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/cffi/cparser.py", line 296, in _parse
    self.convert_pycparser_error(e, csource)
  File "/root/anaconda3/envs/pcc_env/lib/python2.7/site-packages/cffi/cparser.py", line 325, in convert_pycparser_error
    raise CDefError(msg)
cffi.CDefError: cannot parse "int roi_pooling_forward(int pooled_height, int pooled_width, float spatial_scale,"
<cdef source string>:29:82: Illegal character '\r'

Could you help me solve this error ?
Looking forward for your reply, thank you！

Hello!
I found a problem about seg loss in training with my own dataset. My segment datasets were converted to "L". In ori_big.py, model would predict segment with size[x, 2, x, x]. But I got error when training was at CrossEntropyLoss2d. Can you give some help? Thanks!

Error:
/pytorch/aten/src/THCUNN/SpatialClassNLLCriterion.cu:106: cunn_SpatialClassNLLCriterion_updateOutput_kernel: block: [4,0,0], thread: [189,0,0] Assertion t >= 0 && t < n_classes failed.
RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED (createCuDNNHandle at /pytorch/aten/src/ATen/cudnn/Handle.cpp:9)
frame #0: c10::Error::Error(c10::SourceLocation, std::string const&) + 0x46 (0x7f81564a5536 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libc10.so)
frame #1: + 0x10a0c28 (0x7f81579a1c28 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #2: at::native::getCudnnHandle() + 0xe54 (0x7f81579a3404 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #3: + 0xf19f4c (0x7f815781af4c in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #4: + 0xf1afe1 (0x7f815781bfe1 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #5: + 0xf1f01b (0x7f815782001b in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #6: at::native::cudnn_convolution_backward_input(c10::ArrayRef, at::Tensor const&, at::Tensor const&, c10::ArrayRef, c10::ArrayRef, c10::ArrayRef, long, bool, bool) + 0xb2 (0x7f8157820572 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #7: + 0xf86090 (0x7f8157887090 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #8: + 0xfca928 (0x7f81578cb928 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #9: at::native::cudnn_convolution_backward(at::Tensor const&, at::Tensor const&, at::Tensor const&, c10::ArrayRef, c10::ArrayRef, c10::ArrayRef, long, bool, bool, std::array<bool, 2ul>) + 0x4fa (0x7f8157821c0a in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #10: + 0xf863bb (0x7f81578873bb in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #11: + 0xfca984 (0x7f81578cb984 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cuda.so)
frame #12: + 0x2c80736 (0x7f8191037736 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #13: + 0x2ccff44 (0x7f8191086f44 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #14: torch::autograd::generated::CudnnConvolutionBackward::apply(std::vector<at::Tensor, std::allocatorat::Tensor >&&) + 0x378 (0x7f8190c4f908 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #15: + 0x2d89705 (0x7f8191140705 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #16: torch::autograd::Engine::evaluate_function(std::shared_ptrtorch::autograd::GraphTask&, torch::autograd::Node*, torch::autograd::InputBuffer&) + 0x16f3 (0x7f819113da03 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #17: torch::autograd::Engine::thread_main(std::shared_ptrtorch::autograd::GraphTask const&, bool) + 0x3d2 (0x7f819113e7e2 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #18: torch::autograd::Engine::thread_init(int) + 0x39 (0x7f8191136e59 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_cpu.so)
frame #19: torch::autograd::python::PythonEngine::thread_init(int) + 0x38 (0x7f819da7e968 in /home/derek/anaconda3/envs/jim/lib/python3.7/site-packages/torch/lib/libtorch_python.so)
frame #20: + 0xc819d (0x7f81ac9d019d in /home/derek/anaconda3/envs/jim/bin/../lib/libstdc++.so.6)
frame #21: + 0x76db (0x7f81ae1696db in /lib/x86_64-linux-gnu/libpthread.so.0)
frame #22: clone + 0x3f (0x7f81ade9271f in /lib/x86_64-linux-gnu/libc.so.6)

怎么将DULR输出为perspective map (pMap) 的？

where can we download the pretrained model?