Changed addmm_ to new pytorch API to remove deprecated warnings

torchreid/losses/hard_mine_triplet_loss.py

@@ -31,7 +31,7 @@ class TripletLoss(nn.Module):
         # Compute pairwise distance, replace by the official when merged
         dist = torch.pow(inputs, 2).sum(dim=1, keepdim=True).expand(n, n)
         dist = dist + dist.t()
-        dist.addmm_(1, -2, inputs, inputs.t())
+        dist.addmm_(inputs, inputs.t(), beta=1, alpha=-2)
         dist = dist.clamp(min=1e-12).sqrt() # for numerical stability
 
         # For each anchor, find the hardest positive and negative

torchreid/metrics/distance.py

@@ -60,7 +60,7 @@ def euclidean_squared_distance(input1, input2):
     mat1 = torch.pow(input1, 2).sum(dim=1, keepdim=True).expand(m, n)
     mat2 = torch.pow(input2, 2).sum(dim=1, keepdim=True).expand(n, m).t()
     distmat = mat1 + mat2
-    distmat.addmm_(1, -2, input1, input2.t())
+    distmat.addmm_(input1, input2.t(), beta=1, alpha=-2)
     return distmat