Problem for multigpu training about sampler

nerfstudio-project / gsplat

CUDA accelerated rasterization of gaussian splatting

https://docs.gsplat.studio/

Apache License 2.0

1.9k stars 232 forks source link

Problem for multigpu training about sampler #394

Open EhabWilson opened 1 week ago

EhabWilson commented 1 week ago

It is common to use dataloader with distributed sampler when training with multi gpus. So why not use a distributed sampler in examples/simple_trainer.py, is that for any reason?

liruilong940607 commented 1 week ago

I'm not quite familiar with DistributedSampler. What's the benefit of using that?