eric-yyjau/pytorch-superpoint

Issues

• Has anyone reproduced the experimental performance?

  #117 opened 5 months ago
  0

• Is there one less training step?

  #116 opened 7 months ago
  1

• Relationship between superpointNet_gauss2 and superpointNet

  #115 opened 9 months ago
  0

• 两位可以提供一下直接能运行的config吗

  #114 opened 10 months ago
  1

• Hi, What is the formula of sparse loss function?

  #113 opened a year ago
  0

• step 4 : Export/ Evaluate the metrics on HPatches Error:

  #112 opened a year ago
  1

• An error in Training MagicPoint on Synthetic Shapes

  #111 opened a year ago
  6

• When using export.py to extract the coco dataset, there was an issue with the missing loader

  #110 opened a year ago
  4

• #使用网络训练训练生成合成数据集

  #109 opened a year ago
  7

• The original SuperpointNet training don't work

  #108 opened a year ago
  1

• File types for model weights

  #107 opened a year ago
  0

• error：RuntimeError: cuDNN error: CUDNN_STATUS_MAPPING_ERROR

  #105 opened a year ago
  2

• Bugs for euclidean dist

  #104 opened 2 years ago
  0

• Difference between magicpoint_synth_t2 and magicpoint_synth20

  #102 opened 2 years ago
  1

• OSError: [WinError 123] The filename, directory name, or volume label syntax is incorrect

  #101 opened 2 years ago
  2

• Demo for matching a pair of images

  #100 opened 2 years ago
  1

• Results on HPatches between pre-trained model and this Github

  #99 opened 2 years ago
  0

• Training the Original Superpoint Network from scratch

  #98 opened 2 years ago
  0

• How to evaluate the repeatability of detector on my own dataset?

  #97 opened 2 years ago
  0

• If I must change the config all, If I use other dataset?

  #95 opened 2 years ago
  0

• About magicpoint_kitti_train.yaml

  #94 opened 2 years ago
  0

• About training time?

  #93 opened 2 years ago
  2

• About kitti dataset

  #92 opened 2 years ago
  1

• How to export on a custom dataset

  #91 opened 2 years ago
  0

• Training MagicPoint on MS-COCO & how did you design Homographic Adaptation ?

  #90 opened 2 years ago
  9

• CUDA out of memory on 3090

  #89 opened 2 years ago
  4

• error when i use the download lable magicpoint_synth20_homoAdapt100_coco_f1

  #88 opened 2 years ago
  6

• issue: dataloader error

  #87 opened 3 years ago
  1

• No norming of descriptor vectors in loss function?

  #84 opened 3 years ago
  0

• Can I konw the best precition and recall you have gotten?

  #83 opened 3 years ago
  1

• how to train with rgb input?

  #82 opened 3 years ago
  0

• can you release a simple demo，please！

  #81 opened 3 years ago
  0

• Rotation

  #80 opened 3 years ago
  3

• While doing step 4 (exporting) ValueError: could not convert string to float: '0.99653963,-0.016891728,1.6804988'

  #79 opened 3 years ago
  0

• Difference between the models in the repo

  #78 opened 3 years ago
  2

• Look for contributors to maintain this package

  #76 opened 3 years ago
  2

• Magic point training is sluggish.

  #74 opened 3 years ago
  3

• how can i export pretraind model to onnx model?

  #72 opened 3 years ago
  1

• Pixel coordinates from feature extractor

  #71 opened 3 years ago
  1

• superpoint_gauss2

  #70 opened 3 years ago
  2

• Export/ Evaluate the metrics on coco

  #69 opened 3 years ago
  1

• I think there is a bug in subpixel prediction

  #68 opened 3 years ago
  2

• About heatmap size

  #67 opened 3 years ago
  0

• train coco no iteration

  #66 opened 3 years ago
  2

• Nan while joint training (descriptor + detector)

  #65 opened 3 years ago
  0

• Homography of HPatches issue

  #60 opened 3 years ago
  0

• Homographic adapatation time is too long

  #59 opened 3 years ago
  2

• Loader.py error

  #58 opened 3 years ago
  1

• Inference module for custom trained model

  #57 opened 3 years ago
  9

• why need to crop to the same length?

  #55 opened 4 years ago
  2