Thank you for your nice work and impressive results!
I have a question about the positional encoding. In the equation (4) of the paper, input is multiplied by pi before sent into sine and cosine function. But in the run_nerf_helper.py, the class Embedder do not times pi before applying sine and cosine. Did I miss something or the pi does not affect results that much?
Thank you for your nice work and impressive results!
I have a question about the positional encoding. In the equation (4) of the paper, input is multiplied by pi before sent into sine and cosine function. But in the run_nerf_helper.py, the class Embedder do not times pi before applying sine and cosine. Did I miss something or the pi does not affect results that much?