Update Linear_ori.py

Author: Zero-coder

models/Linear_ori.py

@@ -33,22 +33,14 @@
 
 #         self.drop = nn.Dropout(p=0.1)
 #         # Use this line if you want to visualize the weights
-#         # self.Linear.weight = nn.Parameter((1/self.seq_len)*torch.ones([self.pred_len,self.seq_len]))
+#        
 #     def forward(self, x):
 #         # x: [Batch, Input length, Channel]
-#         # x = self.Linear(x.permute(0,2,1))
+#   
 #         x = x.permute(0,2,1) # (B，L,C)=》(B,C,L)
 #         b, c, l = x.size() # (B,C,L)
-#         # y = self.avg_pool(x) # (B,C,L) 通过avg=》 (B,C,1)
-#         # print("y",y.shape)
 #         y = self.avg_pool(x).view(b, c) # (B,C,L) 通过avg=》 (B,C,1)
-#         # print("y",y.shape)
-#         #为了丢给Linear学习，需要view把数据展平开
-#         # y = self.fc(y).view(b, c, 96)
 
-#         y = self.fc(y).view(b,c,1)
-#         # f_weight_np = y.cpu().detach().numpy()
-#         z = self.Linear_More_1(x*y+x)
 
 #         # np.save('f_weight.npy', f_weight_np)
 # #         # np.save('%d f_weight.npy' %epoch, f_weight_np)
@@ -68,9 +60,9 @@ def forward(self, x):
         x_hat = self.layernorm(x)
         bias = torch.mean(x_hat, dim=1).unsqueeze(1).repeat(1, x.shape[1], 1)
         return x_hat - bias
-class Model(nn.Module):#2022.11.7修改前，这个Model能跑通#forMultivariate
+class Model(nn.Module):
 
-    def __init__(self,configs,channel=96,ratio=1):#channel针对ili数据集应该改成36 channel=input_length
+    def __init__(self,configs,channel=96,ratio=1):
         super(Model, self).__init__()
         # self.avg_pool = nn.AdaptiveAvgPool1d(1) #innovation
         self.seq_len = configs.seq_len
@@ -95,7 +87,7 @@ def __init__(self,configs,channel=96,ratio=1):#channel针对ili数据集应该
 
         self.Linear = nn.Linear(self.seq_len, self.pred_len)
         self.Linear_1 = nn.Linear(self.seq_len, self.pred_len)
-        # self.dct_norm = nn.LayerNorm([self.channel_num], eps=1e-6)#作为模块一般normal channel效果好点 for traffic
+        # self.dct_norm = nn.LayerNorm([self.channel_num], eps=1e-6)
         self.dct_norm = nn.LayerNorm(self.seq_len, eps=1e-6)#
         # self.my_layer_norm = nn.LayerNorm([96], eps=1e-6)
     def forward(self, x):
@@ -106,13 +98,11 @@ def forward(self, x):
         b, c, l = x.size() # (B,C,L)
         list = []
 
-        for i in range(c):#i represent channel ，分别对channel的数据做dct
+        for i in range(c):#i represent channel 
             freq=dct.dct(x[:,i,:])     #dct
-            # freq=torch.fft.fft(x[:,i,:])#fft,rfft的输出长度不与输入对齐 #fft结果不能通过linear
-
             # print("freq-shape:",freq.shape)
             list.append(freq)
-            ##把dct结果进行拼接，再进行频率特征学习
+         
 
         stack_dct=torch.stack(list,dim=1) 
         stack_dct = torch.tensor(stack_dct)#(B，L,C)
@@ -121,27 +111,9 @@ def forward(self, x):
         f_weight = self.fc(stack_dct)
         f_weight = self.dct_norm(f_weight)#matters for traffic
 
-        #wo-dct
-        # f_weight = self.dct_norm(x)
-        # f_weight = self.dct_norm(f_weight)#matters for traffic
-        # f_weight = self.fc(f_weight)
-        # f_weight = self.dct_norm(f_weight)#matters for traffic
-        # list_inverse = []
-        # for i in range(c):
-        #     freq_inverse=dct.idct(f_weight[:,i,:])     #dct
-        #     # freq=torch.fft.fft(x[:,i,:])#fft,rfft的输出长度不与输入对齐 #fft结果不能通过linear
-
-        #     # print("freq-shape:",freq.shape)
-        #     list_inverse.append(freq_inverse)
-        #     ##把dct结果进行拼接，再进行频率特征学习
-        # stack_dct_inverse=torch.stack(list_inverse,dim=1) 
-        # stack_dct_inverse = torch.tensor(stack_dct_inverse)#(B，L,C)
-        # # f_weight = self.fc(x)
-        # stack_dct_inverse = self.dct_norm(stack_dct_inverse)#matters for traffic
-        # f_weight_inverse = self.fc_inverse(stack_dct_inverse)
-        # f_weight_inverse = self.dct_norm(f_weight_inverse)#matters for traffic
-
-        #可视化这个频率张量 generalized tensor
+
+
+        #visualization for fecam tensor
         f_weight_cpu = f_weight
 
         f_weight_np = f_weight_cpu.cpu().detach().numpy()
@@ -165,65 +137,12 @@ def forward(self, x):
         # result = self.Linear((x *(f_weight_inverse)))#forL 
         result = self.Linear((x *(f_weight)))#forL 
 
-        # result = result + (1)*torch.mean(result)# for ill 增对数据集而设定的先验知识，可以后续有这方面的思考
+        # result = result + (1)*torch.mean(result)# for ill
         # result_1 = self.Linear_1(x)
         # result = result + result_1
         # result = self.my_layer_norm(result)
 
         return  result.permute(0,2,1)
-        # return  result
-        
-# class Model(nn.Module):#2022.11.7修改前，这个Model能跑通#forMultivariate
-#     def __init__(self,configs,channel=96,ratio=1):#channel针对ili数据集应该改成36 channel=input_length
-#         super(Model, self).__init__()
-#         # self.avg_pool = nn.AdaptiveAvgPool1d(1) #innovation
-#         self.seq_len = configs.seq_len
-#         self.pred_len = configs.pred_len
-#         self.fc = nn.Sequential(
-#                 nn.Linear(channel, channel*4, bias=False),
-#                 nn.Dropout(p=0.1),
-#                 nn.ReLU(inplace=True),
-#                 nn.Linear( channel*4, channel, bias=False),
-#                 nn.Sigmoid()
-#         )
-#         # self.fc_plot = nn.Linear(channel, channel, bias=False)
-#         self.mid_Linear = nn.Linear(self.seq_len, self.seq_len)
-
-#         self.Linear = nn.Linear(self.seq_len, self.pred_len)
-#         self.dct_norm = nn.LayerNorm([7], eps=1e-6)#作为模块一般normal channel效果好点 for traffic
 
-#         # self.my_layer_norm = nn.LayerNorm([96], eps=1e-6)
-#     def forward(self, x):
-#         x = x.permute(0,2,1) # (B，L,C)=》(B,C,L)#forL
-
-        
-#         # x_t = self.mid_Linear(x)
-#         b, c, l = x.size() # (B,C,L)
-#         list = []
-        
-#         for i in range(c):#i represent channel ，分别对channel的数据做dct
-#             freq=dct.dct(x[:,i,:])     #dct
-#             # freq=torch.fft.fft(x[:,i,:])#fft,rfft的输出长度不与输入对齐 #fft结果不能通过linear
-
-#             # print("freq-shape:",freq.shape)
-#             list.append(freq)
-#             ##把dct结果进行拼接，再进行频率特征学习
-        
-        
-      
-#         stack_dct=torch.stack(list,dim=1) 
-#         stack_dct = torch.tensor(stack_dct)#(B，L,C)
-#         f_weight = self.fc(stack_dct)
-
-        
-#         # f_weight = self.dct_norm(f_weight.permute(0,2,1))#matters for traffic
-#         # result = self.Linear(x *(f_weight.permute(0,2,1)))#forL
-
-#         result = self.Linear(x *(f_weight))#forL
-
-        
-#         # result = self.my_layer_norm(result)
-
-#         return  result.permute(0,2,1)