qinzheng93 GeoTransformer issues

qinzheng93 / GeoTransformer

[CVPR2022] Geometric Transformer for Fast and Robust Point Cloud Registration

MIT License

638 stars 67 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

All results are "Nan"

#117 poppo0214 opened 6 days ago
1
Model in training mode during demo.py

#116 braydenrudisill opened 3 weeks ago
0
About results using RANSAC

#115 mooncake199809 closed 1 month ago
1
forward function must have ground transform

#114 adizhol-str opened 1 month ago
0
Are there any tips to improve the performance of GeoTransformer?

#113 zhuozhizhong opened 3 months ago
0
How to select the best checkpoint?

#112 phdymz opened 3 months ago
0
RuntimeError: CUDA error: operation not supported when calling `cusparseCreate(handle)`

#111 Wayne3521 opened 4 months ago
4
Time consumption for model training is a bit abnormal

#110 Baros-Young opened 4 months ago
0
Why is there no change in Registration RR when tested with 2500, 1000, 500, and 250 points respectively.

#109 awordx closed 1 month ago
2
What‘s the difference of the final metrics log between test.py and eval.py???

#108 Wyyspace1203 opened 4 months ago
2
The inference output (estimated_transform) is related to gt_transform?

#107 dspeia closed 4 months ago
1
bad registration result on 3dmatch test data for demo.py

#106 dspeia closed 4 months ago
1
The ground truth does not match the KITTI dataset.

#105 iifeve opened 4 months ago
2
About ModelNet40 Chamfer Distance Result

#104 0xzcy opened 5 months ago
1
Problem with .pth format

#103 Fg9595 opened 5 months ago
1
some problem with KITTI's mean RTE

#102 callmewanzi opened 5 months ago
0
CUDA error: operation not supported when calling `cusparseCreate(handle)`

#101 peachorange124 closed 6 months ago
1
Does gt_transform used during inference?

#100 sss725725 opened 6 months ago
2
registration_with_ransac_from_correspondences

#99 Sucasx closed 6 months ago
0
为什么在训练的第二个epoch时显示cuda out of memory

#98 lemonade-lm opened 7 months ago
11
LINK : fatal error LNK1181: 无法打开输入文件“G:\code\GeoTransformer-main\build\temp.win-amd64-cpython-38\Release\geotransformer\extensions\cpu\grid_subsampling\grid_subsampling.obj”

#97 zsqwe opened 7 months ago
1
How to adjust SuperPoint number to avoid CUDA memory full?

#96 glennliu opened 7 months ago
1
replicate with some problem

#95 Asherchi opened 8 months ago
1
why i test on train data, the result is very bad? and i test on test data, the result is good? i'm confused.

#94 gebinfu opened 8 months ago
0
How can estimated_transform be obtained by src.npy and ref.npy without gt.npy

#93 lotterf opened 8 months ago
1
How can we continue training using the epoch from the previous stage?

#92 17854174130 opened 8 months ago
0
Can not build develop with C++ error

#91 larryaxinc opened 9 months ago
0
How to export the onnx model?

#90 GavinYang5 opened 9 months ago
0
building 'geotransformer.ext' extension :RuntimeError: Error compiling objects for extension

#89 ZhangjLiang opened 9 months ago
0
How were the metadata .pkl files created?

#88 Stelath opened 10 months ago
1
Fix windows build

#87 eclipse0922 opened 10 months ago
0
random_sample_rotation problem

#86 Jianqiuer opened 11 months ago
0
please give me some advice

#85 LeoKingBUAA opened 12 months ago
0
How to visualize the results?

#84 Jayku88 opened 1 year ago
3
When I train with my own data, I encounter an error: RuntimeError: shape '[1, 64, 256]' is invalid for input of size 0.

#83 FDC-WuWeb closed 1 year ago
3
About the setting of hyperparameters

#82 hzhsdie opened 1 year ago
1
Regarding the learning rate issue for multi-GPU training

#81 Wansit99 opened 1 year ago
0
No Module called geotransformer.ext

#80 Pushover-1 closed 10 months ago
2
Question about the results shown on Github

#79 HowLeeHi opened 1 year ago
0
How can I visualize the network model in a detailed manner

#78 yumingjie0405 opened 1 year ago
3
A confused part in the pairwise_distance function

#77 MezereonXP closed 1 year ago
1
can't find the correct file

#76 jeff0908 closed 1 year ago
1
How should I train my dataset?

#75 wyaimyj closed 1 year ago
2
How to operate the model to rotate only around the z-axis instead of in any direction

#74 mcmingchang closed 1 year ago
1
RR is lower than the paper about 1.0%-2.0%

#73 Sabershou opened 1 year ago
2
ModelLoNet results.

#72 Gardlin closed 1 year ago
1
Maybe there is some careless mistake here

#71 PANFEI-CHENG closed 1 year ago
1
About how the results deal with different downsample ratios used in different methods.

#70 chalth opened 1 year ago
1
CUDA out of memory

#69 VladimirYugay opened 1 year ago
6
3DMatch: RRE does not match between val and train

#68 Sigura closed 1 year ago
4