Merge branch 'main' into patch-1

Author: svekars

advanced_source/usb_semisup_learn.py

@@ -81,7 +81,7 @@
 # algorithm on dataset
 # 
 # Note that a CUDA-enabled backend is required for training with the ``semilearn`` package.
-# See `Enabling CUDA in Google Colab <https://pytorch.org/tutorials/beginner/colab#using-cuda>`__ for instructions
+# See `Enabling CUDA in Google Colab <https://pytorch.org/tutorials/beginner/colab#enabling-cuda>`__ for instructions
 # on enabling CUDA in Google Colab.
 #
 import semilearn

beginner_source/knowledge_distillation_tutorial.py

@@ -324,7 +324,7 @@ def train_knowledge_distillation(teacher, student, train_loader, epochs, learnin
             soft_prob = nn.functional.log_softmax(student_logits / T, dim=-1)
 
             # Calculate the soft targets loss. Scaled by T**2 as suggested by the authors of the paper "Distilling the knowledge in a neural network"
-            soft_targets_loss = -torch.sum(soft_targets * soft_prob) / soft_prob.size()[0] * (T**2)
+            soft_targets_loss = torch.sum(soft_targets * (soft_targets.log() - soft_prob)) / soft_prob.size()[0] * (T**2)
 
             # Calculate the true label loss
             label_loss = ce_loss(student_logits, labels)