Ratnesh Madaan

Yes, that is still the case. And is definitely something worth contributing to via a PR. If you need help / feedback on the implementation, let's chat on this thread.

https://github.com/madratman/airsim_standalone/ a bit outdated, but does this help?

well, I did end up upgrading torch, lightning, and cuda as follows: ``` torch: version 1.12.1 pytorch_lightning: version 1.7.3 torch.cuda.version: 11.3 ``` not sure if this is the reason for...

Using the provided environment.yml ``` Traceback (most recent call last): File "/datadrive/projects/bnv_fusion/src/run_e2e.py", line 226, in main pointnet_model.load_state_dict(pretrained_weights["state_dict"]) File "/datadrive/anaconda3/envs/bnv_fusion_orig/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1482, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(...

Notes: - Look into [bingham_mix_t](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L2055) as well. Define the data type. - More methods concerning mixtures: [~~bingham_mixture_sample~~](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L1143) [~~bingham_mixture_pdf~~](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L1480) [bingham_mixture_sample_ridge](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L1494) [bingham_mixture_add](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L1890) [bingham_mixture_mult](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L1942) [bingham_mixture_peak](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L1985) [bingham_mixture_thresh_peaks](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L2002) [bingham_mixture_thresh_weights](https://github.com/SebastianRiedel/bingham/blob/master/c/bingham.c#L2020) Strikethroughs are already there in matlab

This is a bit funny. It doesn't perform as good as the one in C (based on synthetic data for the orientation of a cube in two face). Ideally, both...

Hey, I ll get back to this soon. I stopped using it as I think there's a bug in here. I instead used the C code and wrote a script...

The problem is at [L38-43](https://github.com/SebastianRiedel/bingham/blob/master/matlab/bingham_sample_nd.m#L38-43) in `bingham/matlab/bingham_sample_nd.m` ``` if a > rand() x = x2; p = p2; t = t2; num_accepts = num_accepts + 1; end ``` If a<...

One last bug which I forgot to correct. If you fit a BMM to some data, the last component is (generally) a uniform component, which in the C code has...

An issue is now that the samples coming it this way are a bit noisy