Open sdu-huanhq opened 1 year ago
Hello: I have a little promble: In get_rays( )(nerf_helper.py), you first use torch.meshgrid to create a gird, and then transpose it. But in run_nerf.py you create the coordinates with the following code:
torch.meshgrid
coords = torch.stack( torch.meshgrid( torch.linspace(H//2 - dH, H//2 + dH - 1, 2*dH), torch.linspace(W//2 - dW, W//2 + dW - 1, 2*dW) ), -1)
Why not transpose it before using torch.stack?
torch.stack
Hello: I have a little promble: In get_rays( )(nerf_helper.py), you first use
torch.meshgrid
to create a gird, and then transpose it. But in run_nerf.py you create the coordinates with the following code:Why not transpose it before using
torch.stack
?