[feat] use PosLinear to replace clipper operation

EduCDM/NCDM/NCDM.py

@@ -12,6 +12,12 @@
 from EduCDM import CDM
 
 
+class PosLinear(nn.Linear):
+    def forward(self, input: torch.Tensor) -> torch.Tensor:
+        weight = 2 * F.relu(1 * torch.neg(self.weight)) + self.weight
+        return F.linear(input, weight, self.bias)
+
+
 class Net(nn.Module):
 
     def __init__(self, knowledge_n, exer_n, student_n):
@@ -28,11 +34,11 @@ def __init__(self, knowledge_n, exer_n, student_n):
         self.student_emb = nn.Embedding(self.emb_num, self.stu_dim)
         self.k_difficulty = nn.Embedding(self.exer_n, self.knowledge_dim)
         self.e_difficulty = nn.Embedding(self.exer_n, 1)
-        self.prednet_full1 = nn.Linear(self.prednet_input_len, self.prednet_len1)
+        self.prednet_full1 = PosLinear(self.prednet_input_len, self.prednet_len1)
         self.drop_1 = nn.Dropout(p=0.5)
-        self.prednet_full2 = nn.Linear(self.prednet_len1, self.prednet_len2)
+        self.prednet_full2 = PosLinear(self.prednet_len1, self.prednet_len2)
         self.drop_2 = nn.Dropout(p=0.5)
-        self.prednet_full3 = nn.Linear(self.prednet_len2, 1)
+        self.prednet_full3 = PosLinear(self.prednet_len2, 1)
 
         # initialize
         for name, param in self.named_parameters():
@@ -53,22 +59,6 @@ def forward(self, stu_id, input_exercise, input_knowledge_point):
 
         return output_1.view(-1)
 
-    def apply_clipper(self):
-        clipper = NoneNegClipper()
-        self.prednet_full1.apply(clipper)
-        self.prednet_full2.apply(clipper)
-        self.prednet_full3.apply(clipper)
-
-
-class NoneNegClipper(object):
-    def __init__(self):
-        super(NoneNegClipper, self).__init__()
-
-    def __call__(self, module):
-        if hasattr(module, 'weight'):
-            w = module.weight.data
-            module.weight.data = torch.clamp(w, min=0.).detach()
-
 
 class NCDM(CDM):
     '''Neural Cognitive Diagnosis Model'''
@@ -98,7 +88,6 @@ def train(self, train_data, test_data=None, epoch=10, device="cpu", lr=0.002, si
                 optimizer.zero_grad()
                 loss.backward()
                 optimizer.step()
-                self.ncdm_net.apply_clipper()
 
                 epoch_losses.append(loss.mean().item())