Comments (2)
Hi @Master-cai, thank you for opening this issue!
- It certainly helps a bit in out-of-distribution benchmarks (like HPatches), but less on MegaDepth1500.
- The network could learn to focus more on the confidence prediction than on the matching accuracy, which, in the end, is the major thing we are interested in. This could be solved by sophisticated loss balancing, but we decided to avoid this tedious tuning by not propagating gradients, which we found to be sufficiently accurate as-is.
- We trained the confidence classifier on both tasks, but I expect that just training it on MegaDepth is sufficient. Note that training this classifier is self-supervised, so it can be fine-tuned on any set of image pairs without ground truth labels.
- Indeed, but this is just a difference in notation. We store the non-matchability in the extra row/column. While it is not required during inference, it is required during training (Eq. 11).
- This is indeed not required. We will update the code, thank you for this hint!
- The problem is that bidirectional flash attention does not exist, but running flash attention twice (with shared Q/K) is faster than running bidirectional attention. We tried to implement bidirectional flash attention with Triton but did not succeed in achieving significant speedups until now. However, we do believe that there is a big opportunity for speedups there!
- If I remember correctly this is slightly faster (because of less strided memory accesses in softmax). I agree that they are mathematically equivalent.
from lightglue.
@Phil26AT Thank you for your quick and patient response !
from lightglue.
Related Issues (20)
- can I use this algorithm to get the similarity score of two pictures? HOT 1
- How to get description from giving point(x,y) HOT 2
- The matching effect of similar images is too different HOT 5
- ERROR in the Installation HOT 1
- TSNE on descriptors outputted by SuperPoint?
- How to display the matching status of each pair of feature points HOT 2
- Retrain model LightGlue with ourdataset HOT 2
- While Reproducing the results HOT 1
- Questions about SIFT+LightGlue HOT 3
- spelling error in README HOT 1
- Compatibility Issue with GridSample and ONNX Opset 15 on NXP i.MX 93 NPU HOT 1
- How to do inference with trained model? HOT 1
- No keypoints with sift = error
- FlashAttention actually does not support attention mask HOT 3
- What's the difference between matches0 and matches1
- some problem that delopy on device, such as snapdragon snpe HOT 1
- Training with own dataset on both extractor and matcher HOT 2
- Illegal License HOT 4
- sift+lightglue HOT 5
- LightGlue-full-compile won't work with torch '2.2.1+cu121' HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from lightglue.