Anpei Chen

You can follow the instructions [here](https://github.com/walnut-REE/sof/tree/26473202d5dcd0980bf15b7825e6e9078cca5f56#step-2-optional-inverse-rendering) an inverse rendering of portrait segmaps. For real images, you can parse your portrait images into segmaps with existing face-parsers, but please notice that...

Hi there, thanks, both of these functions can be used for training and rendering, you could use any of them~

Thanks, sorry, I also haven't tested the speed, you can download and try.

Please refer this [function](https://pytorch.org/docs/stable/generated/torch.Tensor.scatter_.html#torch.Tensor.scatter_), thanks.

可能你得检查下pytorch版本

Sounds really reasonable!

Hi, you can follow the instructions in the [NSVF repo](https://github.com/facebookresearch/NSVF#prepare-your-own-dataset), then set the dataset_name to 'tankstemple'.

I have updated the readme, please try again with the instruction :)

Did you mean view-dependent RGB values? That would be a good idea I think, but I am struggling with other things recently:(

Yeah, point cloud sometimes is really useful, actually, you could convert the depth maps and the color result into point could in a simple way from [here](https://github.com/apchenstu/TensoRF/blob/main/renderer.py#L44). You could convert...