nerf-pytorch
nerf-pytorch copied to clipboard
Update run_network() to evaluate in mini-batches
The high-dimensional embedded variable takes a lot of memory. Evaluating it as and when the mini-batch is swapped allows lesser memory usage.