thygate
This script doesn't get enough attention! :) One issue, though, re: 3d PLY generation speed
I tried it for the first time last night and noticed my GPU was running at 0% but my CPU pegged at 100%; it appears to run purely on CPU, which certainly goes a long way to explaining why it's so slow!
Having converted code to PyTorch before to run on GPU, it's actually not as scary as it sounds. torch.Tensors are basically the same thing as numpy.array. There's a couple minor issues that generally come up in the conversion process - for example, they require more explicit conversions in some cases, and a few functions are different - but for the most part, just substitute torch.Tensor as the type for numpy.array. And as a bonus, I also found that for any conversion issues that comes up, ChatGPT usually knows the answer! :)
what do you mean， Im not scriptor, do you mean you changed the code to make it , useingGPU from Mesh to videos?
does it speed up faster?
I tried it for the first time last night and noticed my GPU was running at 0% but my CPU pegged at 100%; it appears to run purely on CPU, which certainly goes a long way to explaining why it's so slow!
Having converted code to PyTorch before to run on GPU, it's actually not as scary as it sounds. torch.Tensors are basically the same thing as numpy.array. There's a couple minor issues that generally come up in the conversion process - for example, they require more explicit conversions in some cases, and a few functions are different - but for the most part, just substitute torch.Tensor as the type for numpy.array. And as a bonus, I also found that for any conversion issues that comes up, ChatGPT usually knows the answer! :)
Are you successful on the GPU? Will it get faster?
I don't understand the question. This was a feature request, not an "I Implemented It Already" :)