Christoph Reich

Remove duplicate entries in i3d_resnet50_v1_sthsthv2.yaml config file.

Hi, just a very small typo fix :)

Nan occurs in backward loss_otherwise

1

Hi, I encounter a weird nan error in general.py during training after multiple epochs. Any idea why this error occurs or how to fix it?  Error message of `torch.autograd.detect_anomaly()`....

Hi @csrhddlam, very nice work! I have a question regarding, using a different resolution at inference. Is there any practical solution to use the Axial-Attention block with a different resolution...

Release of more scenes

5

Hi, thanks for the great work! I would like to know if you plan to offer more rendered scenes in the near future. Especially rendered scenes from your recent Infinigen...

Add option to initialize optical flow in RAFT

1

Hi, this pull request adds support to initialize the optical flow to be refined by the RAFT decoder. The current implementation only supports starting from zero optical flow. However, the...

depth_to_3d_v2 batching Bug

2

### Describe the bug The function [`depth_to_3d_v2`](https://github.com/kornia/kornia/blob/4bd1bd172d27ae0ffb5a811b7338150b65f404dc/kornia/geometry/depth.py#L73) accepts a camera matrix of the shape `(*, 3, 3)`. However, `depth_to_3d_v2` internally uses the [`unproject_meshgrid`](https://github.com/kornia/kornia/blob/4bd1bd172d27ae0ffb5a811b7338150b65f404dc/kornia/geometry/depth.py#L32) fuction that only supports a camera matrix...

Fix device bug in JPEG augmentation

2

#### Changes When running random JPEG augmentation on CUDA, some tensors are not correctly moved to CUDA. To reproduce the bug just run the following code: ```python import kornia import...