Repository Issue Activity (beta)

#187 SuperPoint descriptor map querying using arbitrary pixels

#186 Point pruning + fp16 crashes: torch.zeros missing dtype in the pruning remap

#185 Reproduction results (sp+lg)

#184 Regarding the speed issue when running at a resolution of 1024 * 1024

#183 Raco-aliked weights support only 9 layers

#182 📋 Documentation Enhancement Suggestion

#181 Reproducing MegaDepth1500 pose estimation results

#180 ALIKED performance is slightly worse in new pytorch (CUDA>12.4)

#162 LightGlue is now available in 🤗 Transformers

#177 Can depth maps from different viewpoints be used for extraction and matching?

#130 Differernt SIFT descriptor as input to matcher

#176 Can feature extraction and matching be performed on depth images?

#44 How should I use this project to complete the pose estimation of the target object？

#174 Possibilites for instance level matching

#173 Inquiry about accuracy comparison between 4-heads and 8-heads configuration with 512 keypoints

#171 Batch Inference Code on Transformers

#113 How to do inference with trained model?

#168 torch.amp.custom_fwd Error

#167 train error and KeyError: 'overlap_matrix'

#134 @torch.cuda.amp.custom_fwd deprecated on newer versions of PyTorch

#143 No matches after Training

#161 how to run lightglue with 2 numpy images?

#160 Regarding issues with matching rotated, mirrored, and cropped images

#159 CUDA GPU memory continuously increases when running superpoint algorithm

#158 About dummy keypoints for batch mode