Dmytro Mishkin

What I will test next

33

- Continue random walk on ResNets - to understand how to train them properly. There is definitely somewhere problem I cannot see :( - Pooling: AVG-pooling caffenet Generalizing Pooling Functions...

DynamicSoftMarginLoss

3

https://gfx.cs.princeton.edu/pubs/Zhang_2019_LLD/softmargin.pdf Implemented here, might be possible to borrow with author approval https://github.com/lg-zhang/dynamic-soft-margin-pytorch/blob/master/modules/dynamic_soft_margin.py#L10 

Is there any reason not to include it here? :) Learning Deep Embeddings with Histogram Loss Evgeniya Ustinova, Victor Lempitsky https://arxiv.org/pdf/1611.00822.pdf

#### Changes Adds conversions between graphics and vision coordinate frames. Useful when working with Colmap and others. See https://github.com/colmap/colmap/issues/1184#issuecomment-803003873 and other similar issues. Now people should be able to just...

Loftr does not work with some image size (not a memory issue)

3

### Describe the bug LoFTR incorrectly does something with positional embeddings ``` RuntimeError Traceback (most recent call last) [](https://9t3p2yszpxn-496ff2e9c6d22116-0-colab.googleusercontent.com/outputframe.html?vrz=colab-20220613-060046-RC00_454553376#) in () 10 "image1": torch.rand(1,1, 1704, 2272).cuda()} 11 with torch.no_grad(): --->...

#### Changes Adds AdaLAM w/o refactoring. Plan is the following: - merge https://github.com/kornia/kornia/pull/1813 - add tests here - merge this PR - integrate AdaLAM RANSAC into [kornia.RANSAC](https://kornia.readthedocs.io/en/v0.6.1/geometry.ransac.html) - refactor AdaLAM...

Normalize with intrinsics

4

#### Changes I have refactored out part of project/unproject points to `normalize_with_intrinsics`. The latter function (in their own implementation) is heavily used on OANet and other correspondences filtering methods, e.g....

## 🚀 Feature This is going to be a my personal list of functions we are lacking (or I cannot find them), which I will be adding to kornia when...

List of crashing tests in M1 GPU (mps)

21

### Describe the bug This issue is created to list all the tests cases, which are crashing on M1 GPU (torch.device('mps')), so we can at least run all the tests,...

#### Changes As Pytorch added support for M1 GPU in their recent release candidate, I am adding the option to test on it. Unfortunately, even with gradcheck turned off, there...