fix merge err

slam/algorithms/nice_slam.py

@@ -222,6 +222,8 @@ def get_loss(self,
                  n_iters,
                  coarse=False):
         self.set_stage(is_mapping, step, n_iters, coarse=coarse)
+        if is_mapping:
+            self.model.grid_processing(coarse=coarse)
         model_input = self.get_model_input(optimize_frames, is_mapping)
         model_outputs = self.model(model_input)
         loss_dict = self.model.get_loss_dict(model_outputs, model_input,

slam/model_components/feature_grid_nice.py

@@ -1,9 +1,7 @@
-import numpy as np
 import torch
-import torch.nn as nn
 
 
-class FeatureGrid(nn.Module):
+class FeatureGrid():
     def __init__(self, xyz_len, grid_len, c_dim, std=0.01):
         super(FeatureGrid, self).__init__()
 
@@ -12,16 +10,4 @@ def __init__(self, xyz_len, grid_len, c_dim, std=0.01):
         val_shape = [1, c_dim, *val_shape]
         val = torch.zeros(val_shape).normal_(mean=0, std=std)
 
-        mask = np.ones((val_shape[2:])[::-1]).astype(bool)
-        mask = torch.from_numpy(mask).permute(
-            2, 1, 0).unsqueeze(0).unsqueeze(0).repeat(1, val_shape[1], 1, 1, 1)
-
-        self.val = nn.Parameter(val, requires_grad=True)
-        self.mask = nn.Parameter(mask, requires_grad=False)
-
-    def set_mask(self, new_mask):
-        self.mask.data.copy_(new_mask)
-
-    def val_mask(self):
-        masked_val = self.val * self.mask
-        return masked_val
+        self.val = val