Results copper

Master_Thesis/Current working models/Copper_result.csv

+Reduced input,Ramona norm,method,Kernel,K fold,Seed,Calcium_MAPA,Calcium_R2,Arsenic_MAPA,Arsenic_R2,Lead (Pb+2)_MAPA,Lead (Pb+2)_R2,Copper_MAPA,Copper_R2,Manganese_MAPA,Manganese_R2,Nickel_MAPA,Nickel_R2,Zinc_MAPA,Zinc_R2,Cobalt_MAPA,Cobalt_R2,Aluminium_MAPA,Aluminium_R2,Chromium_MAPA,Chromium_R2,Antimony_MAPA,Antimony_R2,Chloride_MAPA,Chloride_R2,Sulfate_MAPA,Sulfate_R2,Cadmium_MAPA,Cadmium_R2
+False,False,Steady,"['Forward Method:', '                """"""', '                The forward pass of the model.', '    ', '                input: x: torch.Tensor, the input to the model', '    ', '                output: x: torch.Tensor, the output of the model', '                """"""', '                # Amplitude = x[:,index+3].unsqueeze(-1) #Unsqueeze makes it compatible with batch dimensions.', '                # LS = x[:,2].unsqueeze(-1)', '                ', '                ', '                x = self.leaky(self.fc(x)) #Sigmoid early, to get the LS-Element thing going.', '                x = F.elu(-x)', '                x = self.dropout(x)', '                x = self.fcfinal(x)', '                # if torch.isnan(x).any():', '                #     print(""x has a nan value: final "",x)', '                #     quit()', '                weight2 = self.weight2_layer(x)', '                weight = self.weight_layer(x)', '                ', '                x = torch.exp(-x)', '                x = weight*x #This should create an exponential fit like Aexp(-bx)', '                return x + self.trainable_constant #By extension, this should then look like Aexp(-bx) + c', '                # return torch.exp(-x) + self.trainable_constant', '                # return x + self.trainable_constant', '                ', '                ', '                return x #This with lr 0.001 gives 0.74 on first', '', 'Number of epochs: 10000', 'Learning rate: 0.005']",5,42,24123162000000.0,0.8906587219819212
+False,False,Steady,"['Forward Method:', '                """"""', '                The forward pass of the model.', '    ', '                input: x: torch.Tensor, the input to the model', '    ', '                output: x: torch.Tensor, the output of the model', '                """"""', '                # Amplitude = x[:,index+3].unsqueeze(-1) #Unsqueeze makes it compatible with batch dimensions.', '                # LS = x[:,2].unsqueeze(-1)', '                ', '                ', '                x = self.leaky(self.fc(x)) #Sigmoid early, to get the LS-Element thing going.', '                x = F.elu(-x)', '                x = self.dropout(x)', '                x = self.fcfinal(x)', '                # if torch.isnan(x).any():', '                #     print(""x has a nan value: final "",x)', '                #     quit()', '                weight2 = self.weight2_layer(x)', '                weight = self.weight_layer(x)', '                ', '                x = torch.exp(-x)', '                x = weight*x #This should create an exponential fit like Aexp(-bx)', '                return x + self.trainable_constant #By extension, this should then look like Aexp(-bx) + c', '                # return torch.exp(-x) + self.trainable_constant', '                # return x + self.trainable_constant', '                ', '                ', '                return x #This with lr 0.001 gives 0.74 on first', '', 'Number of epochs: 10000', 'Learning rate: 0.005']",5,33,38928780000000.0,0.9225977363384553
+False,False,Steady,"['Forward Method:', '                """"""', '                The forward pass of the model.', '    ', '                input: x: torch.Tensor, the input to the model', '    ', '                output: x: torch.Tensor, the output of the model', '                """"""', '                # Amplitude = x[:,index+3].unsqueeze(-1) #Unsqueeze makes it compatible with batch dimensions.', '                # LS = x[:,2].unsqueeze(-1)', '                ', '                ', '                x = self.leaky(self.fc(x)) #Sigmoid early, to get the LS-Element thing going.', '                x = F.elu(-x)', '                x = self.dropout(x)', '                x = self.fcfinal(x)', '                # if torch.isnan(x).any():', '                #     print(""x has a nan value: final "",x)', '                #     quit()', '                weight2 = self.weight2_layer(x)', '                weight = self.weight_layer(x)', '                ', '                x = torch.exp(-x)', '                x = weight*x #This should create an exponential fit like Aexp(-bx)', '                return x + self.trainable_constant #By extension, this should then look like Aexp(-bx) + c', '                # return torch.exp(-x) + self.trainable_constant', '                # return x + self.trainable_constant', '                ', '                ', '                return x #This with lr 0.001 gives 0.74 on first', '', 'Number of epochs: 10000', 'Learning rate: 0.005']",5,11,10787268000000.0,0.9220868333226546
+False,False,Steady,"['Forward Method:', '                """"""', '                The forward pass of the model.', '    ', '                input: x: torch.Tensor, the input to the model', '    ', '                output: x: torch.Tensor, the output of the model', '                """"""', '                # Amplitude = x[:,index+3].unsqueeze(-1) #Unsqueeze makes it compatible with batch dimensions.', '                # LS = x[:,2].unsqueeze(-1)', '                ', '                ', '                x = self.leaky(self.fc(x)) #Sigmoid early, to get the LS-Element thing going.', '                x = F.elu(-x)', '                x = self.dropout(x)', '                x = self.fcfinal(x)', '                # if torch.isnan(x).any():', '                #     print(""x has a nan value: final "",x)', '                #     quit()', '                weight2 = self.weight2_layer(x)', '                weight = self.weight_layer(x)', '                ', '                x = torch.exp(-x)', '                x = weight*x #This should create an exponential fit like Aexp(-bx)', '                return x + self.trainable_constant #By extension, this should then look like Aexp(-bx) + c', '                # return torch.exp(-x) + self.trainable_constant', '                # return x + self.trainable_constant', '                ', '                ', '                return x #This with lr 0.001 gives 0.74 on first', '', 'Number of epochs: 10000', 'Learning rate: 0.005']",5,5,16684078000000.0,0.8498555899238353
+False,False,Steady,"['Forward Method:', '                """"""', '                The forward pass of the model.', '    ', '                input: x: torch.Tensor, the input to the model', '    ', '                output: x: torch.Tensor, the output of the model', '                """"""', '                # Amplitude = x[:,index+3].unsqueeze(-1) #Unsqueeze makes it compatible with batch dimensions.', '                # LS = x[:,2].unsqueeze(-1)', '                ', '                ', '                x = self.leaky(self.fc(x)) #Sigmoid early, to get the LS-Element thing going.', '                x = F.elu(-x)', '                x = self.dropout(x)', '                x = self.fcfinal(x)', '                # if torch.isnan(x).any():', '                #     print(""x has a nan value: final "",x)', '                #     quit()', '                weight2 = self.weight2_layer(x)', '                weight = self.weight_layer(x)', '                ', '                x = torch.exp(-x)', '                x = weight*x #This should create an exponential fit like Aexp(-bx)', '                return x + self.trainable_constant #By extension, this should then look like Aexp(-bx) + c', '                # return torch.exp(-x) + self.trainable_constant', '                # return x + self.trainable_constant', '                ', '                ', '                return x #This with lr 0.001 gives 0.74 on first', '', 'Number of epochs: 10000', 'Learning rate: 0.005']",5,2,75059000000000.0,0.9182668391462008