add support for unlabel training (#2103)

ppcls/loss/distillationloss.py

@@ -236,8 +236,13 @@ class DistillationDKDLoss(DKDLoss):
                  temperature=1.0,
                  alpha=1.0,
                  beta=1.0,
+                 use_target_as_gt=False,
                  name="loss_dkd"):
-        super().__init__(temperature=temperature, alpha=alpha, beta=beta)
+        super().__init__(
+            temperature=temperature,
+            alpha=alpha,
+            beta=beta,
+            use_target_as_gt=use_target_as_gt)
         self.key = key
         self.model_name_pairs = model_name_pairs
         self.name = name

ppcls/loss/dkdloss.py

@@ -10,13 +10,20 @@ class DKDLoss(nn.Layer):
     Code was heavily based on https://github.com/megvii-research/mdistiller
     """
 
-    def __init__(self, temperature=1.0, alpha=1.0, beta=1.0):
+    def __init__(self,
+                 temperature=1.0,
+                 alpha=1.0,
+                 beta=1.0,
+                 use_target_as_gt=False):
         super().__init__()
         self.temperature = temperature
         self.alpha = alpha
         self.beta = beta
+        self.use_target_as_gt = use_target_as_gt
 
-    def forward(self, logits_student, logits_teacher, target):
+    def forward(self, logits_student, logits_teacher, target=None):
+        if target is None or self.use_target_as_gt:
+            target = logits_teacher.argmax(axis=-1)
         gt_mask = _get_gt_mask(logits_student, target)
         other_mask = 1 - gt_mask
         pred_student = F.softmax(logits_student / self.temperature, axis=1)