What is the overall reasoning speed? Are there relevant test data?

Parskatt / RoMa

[CVPR 2024] RoMa: Robust Dense Feature Matching; RoMa is the robust dense feature matcher capable of estimating pixel-dense warps and reliable certainties for almost any image pair.

https://parskatt.github.io/RoMa/

MIT License

630 stars 51 forks source link

What is the overall reasoning speed? Are there relevant test data? #42

Open tianyi-1349 opened 5 months ago

tianyi-1349 commented 5 months ago

What is the overall reasoning speed? Are there relevant test data? tank you.

Parskatt commented 3 weeks ago

Sorry for the late reply. Could you clarify the second question?

Regarding the first question, it depends somewhat on the output resolution and batch size. For reproducing e.g. our mega1500 test results, the inference speed is roughly ~1 second per image pair. If you run in batched mode on lower resolution you might get down to about 0.1 seconds.