diff --git a/codes/models/archs/ChainedEmbeddingGen.py b/codes/models/archs/ChainedEmbeddingGen.py
index d82f00d7..4e454a27 100644
--- a/codes/models/archs/ChainedEmbeddingGen.py
+++ b/codes/models/archs/ChainedEmbeddingGen.py
@@ -71,6 +71,8 @@ class ChainedEmbeddingGenWithStructure(nn.Module):
         self.recurrent = recurrent
         self.initial_conv = ConvGnLelu(3, 64, kernel_size=7, bias=True, norm=False, activation=False)
         if recurrent:
+            self.recurrent_nf = recurrent_nf
+            self.recurrent_stride = recurrent_stride
             self.recurrent_process = ConvGnLelu(recurrent_nf, 64, kernel_size=3, stride=recurrent_stride, norm=False, bias=True, activation=False)
             self.recurrent_join = ReferenceJoinBlock(64, residual_weight_init_factor=.01, final_norm=False, kernel_size=1, depth=3, join=False)
         self.spine = SpineNet(arch='49', output_level=[3, 4], double_reduce_early=False)
@@ -86,7 +88,12 @@ class ChainedEmbeddingGenWithStructure(nn.Module):
         fea = self.initial_conv(x)
         if self.recurrent:
             if recurrent is None:
-                recurrent = torch.zeros_like(fea)
+                if self.recurrent_nf == 3:
+                    recurrent = torch.zeros_like(x)
+                    if self.recurrent_stride != 1:
+                        recurrent = torch.nn.functional.interpolate(recurrent, scale_factor=self.recurrent_stride, mode='nearest')
+                else:
+                    recurrent = torch.zeros_like(fea)
             rec = self.recurrent_process(recurrent)
             fea, recstd = self.recurrent_join(fea, rec)
             self.ref_join_std = recstd.item()
@@ -101,4 +108,4 @@ class ChainedEmbeddingGenWithStructure(nn.Module):
         return out, self.grad_extract(checkpoint(self.structure_upsample, grad)), self.grad_extract(out), fea
 
     def get_debug_values(self, step, net_name):
-        return { 'ref_join_std': self.ref_join_std }
\ No newline at end of file
+        return { 'ref_join_std': self.ref_join_std }