Update show.

tools/show.py

@@ -78,11 +78,12 @@ def DumpTensorToImage(tensor, name, forceSquare=False, scale=1.0, Contrast=None,
 
 
 def DumpTensorToLog(tensor, name="log"):
-    shape = tensor.shape
     tensor_mean = torch.mean(tensor).cpu().detach().numpy()
+    tensor_abs_mean = torch.mean(torch.abs(tensor)).cpu().detach().numpy()
     tensor_range = (torch.max(tensor) - torch.min(tensor)).cpu().detach().numpy()
     f = open(name, "w")
     f.writelines("tensor mean: %s" % tensor_mean + os.linesep)
+    f.writelines("tensor abs mean: %s" % tensor_abs_mean + os.linesep)
     f.writelines("tensor range: %s" % tensor_range + os.linesep)
     data = tensor.reshape([-1]).float().cpu().detach().numpy().tolist()
     for d in data:

unsuper/minist.py

@@ -20,8 +20,8 @@ np.random.seed(seed)
 random.seed(seed)
 
 
-device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-# device = torch.device("cpu")
+# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+device = torch.device("cpu")
 # device = torch.device("mps")
 
 num_epochs = 1
@@ -128,8 +128,8 @@ for epoch in range(epochs):
         all = range(0, sample.shape[0])
         # ratio_max = abs / mean
         # ratio_nor = (max - abs) / max
-        ratio_nor = torch.pow(abs / mean, 4)
         # ratio_nor[all, max_index] = ratio_max[all, max_index].clone()
+        ratio_nor = torch.pow(abs / mean, 4)
         ratio_nor = torch.where(torch.isnan(ratio_nor), 1.0, ratio_nor)
         label = sample * ratio_nor
 
@@ -197,7 +197,6 @@ with torch.no_grad():
         labels = labels.to(device)
         outputs = model(images)
 
-        # max returns (value ,index)
         _, predicted = torch.max(outputs.data, 1)
         n_samples += labels.size(0)
         n_correct += (predicted == labels).sum().item()