princeton-vl/CornerNet-Lite

Issues

• The speed of training is slower when I use more gpus

  #110 opened 5 years ago
  4

• How can I train on my own dataset to detect new classes?

  #109 opened 5 years ago
  1

• _ZN2at6detail20DynamicCUDAInterface10set_deviceE

  #108 opened 5 years ago
  1

• cuda10 train error

  #107 opened 5 years ago
  1

• how to calculate mean, std eigenvalue and eigenvector?

  #106 opened 5 years ago
  3

• How to detect only one class and output the result

  #105 opened 5 years ago
  3

• why? How? train error 0%| | 0/100000 [00:00<?, ?it/s]

  #104 opened 5 years ago
  6

• Can this model be invoked through the c++ interface?

  #103 opened 5 years ago
  0

• questions about the output dims of heatmaps

  #102 opened 5 years ago
  2

• Inference on Video is really SLOW! Helps needed

  #101 opened 5 years ago
  1

• 程序中断的话，train函数里的线程会自动停止吗？

  #100 opened 5 years ago
  0

• What should the custom dataset's format should look like?

  #99 opened 5 years ago
  1

• RuntimeError: cannot join current thread

  #98 opened 5 years ago
  3

• how to solve the issue "None type has attribute shape"

  #97 opened 5 years ago
  4

• There is an error when i traiing.

  #96 opened 5 years ago
  4

• Where do the targets for CornerNet_Loss() come from?

  #95 opened 5 years ago
  1

• The detection effect is much worse than yolov3.

  #94 opened 5 years ago
  5

• torch.cuda.is_available() is False

  #93 opened 5 years ago
  1

• result on iou>0.5?

  #92 opened 5 years ago
  1

• How to prepare the input file if I want to train it with ava dataset?

  #91 opened 5 years ago
  0

• how to convert pascal xml to coco json using python?

  #90 opened 5 years ago
  1

• Specify which GPU to use for DataParallel

  #89 opened 5 years ago
  1

• Hi,I have this problem in training the CornerNet!

  #88 opened 5 years ago
  1

• RuntimeError: CUDA error: all CUDA-capable devices are busy or unavailable

  #87 opened 5 years ago
  0

• Conver model to onnx error

  #86 opened 5 years ago
  1

• train cornernet-squeeze error

  #85 opened 3 years ago
  8

• Can it work for Voc datasets?

  #84 opened 5 years ago
  2

• Problem of use pretrained model, when i train own dataset

  #83 opened 6 years ago
  4

• Won't the step of random crop around a bbox cause any problem?

  #82 opened 6 years ago
  0

• 为什么我测试的几张图片，检测结果和时间都比yolov3-tiny差很多？

  #81 opened 6 years ago
  1

• When I have two locating kps, when I test my own images?

  #80 opened 5 years ago
  0

• Segmentation fault (core dumped)

  #79 opened 6 years ago
  5

• the result of the test not so good, many targets are not detected?，use the CornerNet_Saccade_500000.pkl

  #78 opened 6 years ago
  2

• Should training suppercategory id and validation supercategory id be the same while training?

  #77 opened 6 years ago
  1

• wrong with test

  #76 opened 6 years ago
  7

• how to train and test my data

  #75 opened 6 years ago
  6

• rm

  #74 opened 6 years ago
  0

• about course bbox prediction

  #73 opened 6 years ago
  0

• Could you release the training log of CornerNet_squeeze ?

  #72 opened 6 years ago
  0

• Device index must be -1 or non-negative, got -1160

  #71 opened 6 years ago
  4

• WIN10下已经解决了安装和训练，还有问题的，可以加Q群264191384 注明（CornerNet-Lite）共同研究，谢谢

  #70 opened 6 years ago
  1

• undefined symbol: _ZTIN3c1021AutogradMetaInterfaceE".that's why?

  #69 opened 6 years ago
  9

• src/top_pool.cpp:52:23: error: ‘zeros’ is not a member of ‘torch’ auto gt_mask = torch::zeros({batch, channel, width}, at::device(at::kCUDA).dtype(at::kByte));

  #68 opened 6 years ago
  2

• can't find the dir "<CornerNet-Lite dir>/cache/nnet/CornerNet_Saccade/`,"

  #67 opened 6 years ago
  1

• command 'gcc' failed with exit status 1

  #66 opened 6 years ago
  5

• How to train the CornerNet_Saccade in a distributed way

  #65 opened 6 years ago
  3

• When I generated the Pt file, I got a error

  #64 opened 6 years ago
  0

• When I generated the Pt file, I got a error

  #63 opened 6 years ago
  3

• FPS is too slow?

  #62 opened 6 years ago
  4

• FPS is too low?

  #61 opened 6 years ago
  0