update readme

nerf/renderer.py

@@ -320,6 +320,12 @@ class NeRFRenderer(nn.Module):
         nears.unsqueeze_(-1)
         fars.unsqueeze_(-1)
 
+        # random sample light_d if not provided
+        if light_d is None:
+            # gaussian noise around the ray origin, so the light always face the view dir (avoid dark face)
+            light_d = - (rays_o[0] + torch.randn(3, device=device, dtype=torch.float))
+            light_d = safe_normalize(light_d)
+
         #print(f'nears = {nears.min().item()} ~ {nears.max().item()}, fars = {fars.min().item()} ~ {fars.max().item()}')
 
         z_vals = torch.linspace(0.0, 1.0, num_steps, device=device).unsqueeze(0) # [1, T]

readme.md

@@ -104,13 +104,13 @@ latents.backward(gradient=grad, retain_graph=True)
 * Other regularizations are in `./nerf/utils.py > Trainer > train_step`. 
     * The generation seems quite sensitive to regularizations on weights_sum (alphas for each ray). The original opacity loss tends to make NeRF disappear (zero density everywhere), so we use an entropy loss to replace it for now (encourages alpha to be either 0 or 1).
 * NeRF Rendering core function: `./nerf/renderer.py > NeRFRenderer > run_cuda`.
+    * the occupancy grid based training acceleration (instant-ngp like, enabled by `--cuda_ray`) may harm the generation progress, since once a grid cell is marked as empty, rays won't pass it later...
 * Shading & normal evaluation: `./nerf/network*.py > NeRFNetwork > forward`. Current implementation harms training and is disabled.
-    * use `--albedo_iters 1000` to enable random shading mode after 1000 steps from albedo, lambertian ,and textureless
+    * use `--albedo_iters 1000` to enable random shading mode after 1000 steps from albedo, lambertian, and textureless.
     * light direction: current implementation use a plane light source, instead of a point light source...
 * View-dependent prompting: `./nerf/provider.py > get_view_direction`.
     * ues `--angle_overhead, --angle_front` to set the border. How to better divide front/back/side regions?
 * Network backbone (`./nerf/network*.py`) can be chosen by the `--backbone` option, but `tcnn` and `vanilla` are not well tested.
-    * the occupancy grid based training acceleration (instant-ngp like) may harm the generation progress, since once a grid cell is marked as empty, rays won't pass it later. 
 * Spatial density bias (gaussian density blob): `./nerf/network*.py > NeRFNetwork > gaussian`.
 
 # Acknowledgement