refine code

applications/tools/first-order-demo.py

@@ -3,10 +3,6 @@ matplotlib.use('Agg')
 import os
 import sys
 
-# cur_path = os.path.abspath(os.path.dirname(__file__))
-# root_path = os.path.split(cur_path)[0]
-# sys.path.append(root_path)
-
 import yaml
 import pickle
 from argparse import ArgumentParser

ppgan/models/generators/occlusion_aware.py

@@ -81,9 +81,10 @@ class OcclusionAwareGenerator(nn.Layer):
             deformation = deformation.transpose([0, 3, 1, 2])
             deformation = F.interpolate(deformation,
                                         size=(h, w),
-                                        mode='bilinear')
+                                        mode='bilinear',
+                                        align_corners=False)
             deformation = deformation.transpose([0, 2, 3, 1])
-        return F.grid_sample(inp, deformation)
+        return F.grid_sample(inp, deformation, align_corners=False)
 
     def forward(self, source_image, kp_driving, kp_source):
         # Encoding (downsampling) part
@@ -113,7 +114,8 @@ class OcclusionAwareGenerator(nn.Layer):
                         3] != occlusion_map.shape[3]:
                     occlusion_map = F.interpolate(occlusion_map,
                                                   size=out.shape[2:],
-                                                  mode='bilinear')
+                                                  mode='bilinear',
+                                                  align_corners=False)
                 out = out * occlusion_map
 
             output_dict["deformed"] = self.deform_input(source_image,

ppgan/modules/first_order.py

@@ -2,8 +2,6 @@ import paddle
 import paddle.nn as nn
 import paddle.nn.functional as F
 
-# from sync_batchnorm import SynchronizedBatchNorm2d as BatchNorm2d
-
 
 def kp2gaussian(kp, spatial_size, kp_variance):
     """
@@ -44,7 +42,6 @@ def make_coordinate_grid(spatial_size, type):
     xx = paddle.tile(x.reshape([1, -1]), [h, 1])
 
     meshed = paddle.concat([xx.unsqueeze(2), yy.unsqueeze(2)], 2)
-    # meshed = paddle.concat([xx.unsqueeze_(2), yy.unsqueeze_(2)], 2)
 
     return meshed
 
@@ -261,10 +258,7 @@ class AntiAliasInterpolation2d(nn.Layer):
         # Make sure sum of values in gaussian kernel equals 1.
         kernel = kernel / paddle.sum(kernel)
         # Reshape to depthwise convolutional weight
-        # print('debug shape:', kernel.shape)
-        # print('debug shape 1:', kernel.dim())
         kernel = kernel.reshape([1, 1, *kernel.shape])
-        # kernel = kernel.repeat(channels, *[1] * (kernel.dim() - 1))
         kernel = paddle.tile(kernel, [channels, *[1] * (kernel.dim() - 1)])
 
         self.register_buffer('weight', kernel)

requirments.txt

-tqdm
+tqdm
\ No newline at end of file
+PyYAML>=5.1
+scikit-image>=0.14.0
+scipy>=1.1.0