Nanne/pytorch-NetVlad

Issues

• np.square(knn.kneighbors(clsts, 2)[1]) Should be changed?

  #84 opened 2 years ago by wpumain
  3

• Can't get Pittsburgh 250k dataset, link failed

  #69 opened 3 years ago by Nothingzl
  1

• No found image in pitts250k

  #93 opened 6 months ago by Junshuai1024
  1

• Is it possible to convert to ONNX format?

  #88 opened 6 months ago by fettahyildizz
  1

• There is no further progress in the code

  #89 opened 6 months ago by HyeokJaeKiM
  1

• Question about the number of Negatives during training process.

  #92 opened a year ago by euncheolChoi
  2

• tokyo24/7 dataset training

  #68 opened 3 years ago by TuShuqi
  9

• recall

  #91 opened a year ago by whuzs
  1

• Where C_ks are getting updated in the code

  #90 opened a year ago by Anuradha-Uggi
  3

• Converting MATLAB trained model .mat files (off the shelf and trained on Pittsburg 30k) in Pytorch

  #65 opened a year ago by SharangKaul123
  0

• how to train network with tokyoTM?

  #79 opened a year ago by LKELN
  0

• Running "python main.py --mode=train --arch=vgg16 --pooling=netvlad --num_clusters=64" leads to this error. Help?

  #86 opened 2 years ago by taiyipan
  3

• What is the mathematical formula on which these lines of code are based?

  #83 opened 2 years ago by wpumain
  1

• How can I get output image in code?

  #85 opened 2 years ago by ans2568
  0

• Should the position of this code be adjusted?

  #82 opened 2 years ago by wpumain
  5

• There is no specific data of the image in the【pitts30k_train.mat】 file?

  #81 opened 2 years ago by wpumain
  1

• Which version of pytorch is the highest supported by this program?

  #80 opened 2 years ago by wpumain
  1

• how to

  #78 opened 2 years ago by LKELN
  0

• CUDA out of memory

  #77 opened 2 years ago by Anuradha-Uggi
  5

• using in remote sensing

  #76 opened 2 years ago by rafal7466
  1

• How to fix "TMPDIR" error?

  #75 opened 2 years ago by Doublewater-jpg
  1

• Recall is always calculated as 1.

  #74 opened 2 years ago by JSP-ywu
  1

• How to implement this code to another dataset?

  #73 opened 2 years ago by JSP-ywu
  4

• how to use tokyo 24/7 train Netvlad

  #67 opened 3 years ago by Durexer
  3

• the training time

  #72 opened 3 years ago by KN-Zhang
  1

• why we need to initialise the NetVlad layer

  #71 opened 3 years ago by zhenjiangw
  1

• get_clusters function: nDescriptors and nPerImage

  #70 opened 3 years ago by SharangKaul123
  1

• Question about the positives and negatives

  #66 opened 3 years ago by salt0107fish
  2

• Input type and weight type are not the same

  #62 opened 3 years ago by TuShuqi
  4

• \pytorch-NetVlad-master\pittsburgh.py", line 188, in collate_fn query, positive, negatives, indices = zip(*batch) ValueError: not enough values to unpack (expected 4, got 2)

  #64 opened 3 years ago by pareespathak
  1

• FileNotFoundError: [Errno 2] No such file or directory:000/000000_pitch1_yaw1.jpg

  #63 opened 3 years ago by pareespathak
  1

• TypeError: Indexing arrays must have integer dtypes occured when training

  #61 opened 3 years ago by 2u4dk123
  10

• Unexpected key(s) in state_dict: "pool.centroids", "pool.conv.weight".

  #54 opened 4 years ago by DonMuv
  1

• Faiss k-means clustering error while calculating the optimal cluster centroids

  #60 opened 3 years ago by SharangKaul123
  1

• Question about the training dataset

  #59 opened 3 years ago by hyeongilee
  1

• Pittsburgh 250k dataset seems not avaliable now

  #58 opened 3 years ago by ZZQzzq
  1

• sqrt() of the margin in Triplet loss

  #57 opened 4 years ago by dinarkino
  2

• cannot test 85.2

  #56 opened 4 years ago by DonMuv
  1

• What's the difference between NetVLAD V1 and V2?

  #55 opened 4 years ago by double-fire-0
  1

• About test on other datasets

  #53 opened 4 years ago by bo1234566
  1

• Questions about the main.py code.

  #52 opened 4 years ago by JeonHyeongJunKW
  1

• When I train the model, the following is printed out.

  #51 opened 4 years ago by JeonHyeongJunKW
  3

• Training on custom dataset

  #47 opened 4 years ago by Zumbalamambo
  0

• terminology clarification

  #50 opened 4 years ago by mrgransky
  1

• Userwarning: loky-backed parallel loops cannot be called in a multiprocessing, setting n_jobs=1

  #49 opened 4 years ago by mrgransky
  3

• Error with faiss kmeans argument

  #48 opened 4 years ago by mrgransky
  2

• License

  #46 opened 4 years ago by Tobias-Fischer
  2

• Sorry for another question!

  #45 opened 4 years ago by Fleurrr
  1

• Error with Faiss!

  #44 opened 4 years ago by Fleurrr
  2

• image normalization

  #43 opened 4 years ago by Anonymous-so
  1