apchenstu/mvsnerf

Issues

• Each run reports a different error： No such file or directory: './dataset/dtu/Rectified/scan1_train/rect_011_3_r5000.png'

  #88 opened a year ago by HuaandQi
  2

• how to train non per scene optimization mvsnerf?

  #96 opened 8 months ago by pnpmpnp
  0

• run training example

  #94 opened 8 months ago by 19991105
  0

• Depth resolution

  #95 opened a year ago by hanxuel
  0

• DTU Training for NeRF

  #55 opened 3 years ago by chaupham1709
  2

• Save evaluation results?

  #59 opened 2 years ago by sergeyprokudin
  3

• About the function get_ndc_coordinate()

  #67 opened 2 years ago by Qi-Zhangyang
  1

• Depth estimation

  #93 opened a year ago by VLadImirluren
  0

• Why finetuning MVSNet

  #75 opened 2 years ago by ldz666666
  3

• When train llff, keyerror 'w2cs' occurs in validation step

  #92 opened a year ago by everythoughthelps
  0

• The DTU file tree for fine-tune

  #61 opened 2 years ago by YuhsiHu
  4

• The relationship between DTU training data and Depth_raw

  #82 opened 2 years ago by kongdy1999
  2

• Pre-trained checkpoint

  #91 opened a year ago by basit-7
  0

• win_size exceeds image extent. Either ensure that your images are at least 7x7;

  #90 opened a year ago by Shichao123123
  0

• How can I get the new depth map

  #89 opened a year ago by Gxhappiness
  0

• How many views are used during finetuning?

  #87 opened a year ago by pansanity666
  0

• ValueError: The provided lr scheduler "<torch.optim.lr_scheduler.MultiStepLR object at 0x000001FE8239EAF0>" is invalid

  #84 opened 2 years ago by X86Git
  3

• Test the DTU dataset with IBRNet

  #86 opened a year ago by nihao-eng
  0

• TypeError: cannot unpack non-iterable NoneType object

  #85 opened 2 years ago by Shichao123123
  1

• RuntimeError: Function 'PowBackward0' returned nan values in its 0th output.

  #83 opened 2 years ago by Hitouo
  0

• Running on Google Colab and a CKPT bug

  #81 opened 2 years ago by ichxorya
  0

• can't run training command example

  #77 opened 2 years ago by yfq512
  3

• is the ckpt in the directory of ckpts the best of the result

  #80 opened 2 years ago by jssyzsfzy
  1

• Some problems on rendering dtu

  #79 opened 2 years ago by Hitouo
  1

• why are you not using dir_embedder in the code but writing in the paper ?

  #78 opened 2 years ago by 18128381510
  0

• On extrinsic preprocess

  #76 opened 2 years ago by ldz666666
  1

• About less than 3 images to reconstruct mesh

  #74 opened 2 years ago by zaichenhu
  0

• Cannot reproduce the results on LLFF dataset

  #69 opened 2 years ago by cwchenwang
  3

• DTU evaluation accuracy?

  #48 opened 3 years ago by hdjang
  7

• It seems that you are not using dir_embedder

  #73 opened 2 years ago by caiyongqi
  0

• About your volume rendering implementation

  #72 opened 2 years ago by jcliu0428
  0

• Question about the position input of the MLP x.

  #71 opened 2 years ago by GostInShell
  0

• About batchsize

  #68 opened 2 years ago by Qi-Zhangyang
  0

• target view problem in training

  #66 opened 2 years ago by Tuich
  1

• CUDA error

  #60 opened 2 years ago by YuhsiHu
  2

• Is it necessary to load volume in ckpt?

  #64 opened 2 years ago by zinsama
  0

• RuntimeError: CUDA out of memory.

  #58 opened 3 years ago by zhichengLuxx
  2

• Determine how generalizable he is?

  #63 opened 2 years ago by AIJIA000
  0

• About the depth_loss, if using it, will the result get better or worse ?

  #62 opened 2 years ago by MaxChanger
  0

• Why doesn't rays_d divide his norm before converting to NDC coordinates?

  #57 opened 3 years ago by MELANCHOLY828
  1

• finetuning results are much worse than general results, why??

  #46 opened 3 years ago by hengfei-wang
  4

• few views in training

  #54 opened 3 years ago by caiyongqi
  2

• UV Coordinate

  #56 opened 3 years ago by caiyongqi
  0

• H*W or W*H?

  #53 opened 3 years ago by caiyongqi
  2

• How to reproduce results?

  #52 opened 3 years ago by derrick-xwp
  2

• Resolution of images in training

  #50 opened 3 years ago by hengfei-wang
  1

• the meaning of unpreprocess function

  #49 opened 3 years ago by c03n424
  1

• RuntimeError: Tried to access nonexistent attribute or method '__call__' of type '__torch__.kornia.geometry.boxes.Boxes3D'.

  #51 opened 3 years ago by derrick-xwp
  0

• [Question] Is it able to train with images of texturelss objects?

  #45 opened 3 years ago by dedoogong
  0

• Why using the training mode of MVSNet during the validation?

  #44 opened 3 years ago by ewrfcas
  0