Add DataParallel support on new codebase

2 files changed, 8 insertions, 1 deletions

diff --git a/models/model.py b/models/model.py
index f79b832..532ded3 100644
--- a/models/model.py
+++ b/models/model.py
@@ -150,6 +150,7 @@ class Model:
         self.rgb_pn = RGBPartNet(self.in_channels, **model_hp,
                                  image_log_on=self.image_log_on)
         # Try to accelerate computation using CUDA or others
+        self.rgb_pn = nn.DataParallel(self.rgb_pn)
         self.rgb_pn = self.rgb_pn.to(self.device)
         self.optimizer = optim.Adam([
             {'params': self.rgb_pn.ae.parameters(), **ae_optim_hp},
@@ -196,6 +197,12 @@ class Model:
             # Duplicate labels for each part
             y = y.unsqueeze(1).repeat(1, self.rgb_pn.num_total_parts)
             losses, images = self.rgb_pn(x_c1, x_c2, y)
+            losses = torch.stack((
+                # xrecon           cano_cons         pose_sim
+                losses[0].sum(), losses[1].mean(), losses[2].mean(),
+                # hpm_ba_trip       pn_ba_trip
+                losses[3].mean(), losses[4].mean()
+            ))
             loss = losses.sum()
             loss.backward()
             self.optimizer.step()
diff --git a/models/rgb_part_net.py b/models/rgb_part_net.py
index 2aa680c..66609fd 100644
--- a/models/rgb_part_net.py
+++ b/models/rgb_part_net.py
@@ -83,7 +83,7 @@ class RGBPartNet(nn.Module):
             pn_ba_trip = self.pn_ba_trip(
                 x[self.hpm_num_parts:], y[self.hpm_num_parts:]
             )
-            losses = torch.stack((*losses, hpm_ba_trip, pn_ba_trip))
+            losses = (*losses, hpm_ba_trip, pn_ba_trip)
             return losses, images
         else:
             return x.unsqueeze(1).view(-1)