add minactive return in search

.gitignore

@@ -7,3 +7,4 @@ Dataset/
 .vscode
 /*/__pycache__
 .mypy_cache
+/FilterEvaluator/image*

FilterEvaluator/Evaluator.py

@@ -51,7 +51,7 @@ layer = 0
 # traindata, testdata = Loader.RandomMnist(batchsize, style="VerticalOneLine")
 # traindata, testdata = Loader.RandomMnist(batchsize, style="VerticalZebra")
 # traindata, testdata = Loader.Cifar10Mono(batchsize)
-traindata, testdata = Loader.Cifar10Mono(batchsize, num_workers=0, shuffle=True)
+traindata, testdata = Loader.Cifar10Mono(batchsize, num_workers=0, shuffle=False)
 
 
 
@@ -71,17 +71,31 @@ traindata, testdata = Loader.Cifar10Mono(batchsize, num_workers=0, shuffle=True)
 
 
 
+for batch_idx, (data, target) in enumerate(traindata):
+    utils.NumpyToImage(data.cpu().detach().numpy(), CurrentPath+"image", title="TrainData")
+    break
 
-
+weight,active = EvaluatorUnsuper.UnsuperLearnSearchWeight(model, layer, traindata, NumSearch=1,SaveChannel=8,SearchChannelRatio=1, Interation=128)
-# weight = EvaluatorUnsuper.UnsuperLearnSearchWeight(model, layer, traindata, NumSearch=100000, SearchChannelRatio=32, Interation=10)
-# np.save("WeightSearch.npy", weight)
-weight = np.load(CurrentPath+"WeightSearch.npy")
 utils.NumpyToImage(weight, CurrentPath+"image",title="SearchWeight")
+
+b =0
+
+
+
+
+
+
+
+
+# weight,active = EvaluatorUnsuper.UnsuperLearnSearchWeight(model, layer, traindata, NumSearch=100000, SearchChannelRatio=32, Interation=10)
+# np.save("WeightSearch.npy", weight)
+# weight = np.load(CurrentPath+"WeightSearch.npy")
+# utils.NumpyToImage(weight, CurrentPath+"image",title="SearchWeight")
 # weight = np.load(CurrentPath+"WeightSearch.npy")
 # bestweight,index = EvaluatorUnsuper.UnsuperLearnFindBestWeight(model,layer,weight,traindata,128,100000)
 # np.save(CurrentPath+"bestweightSearch.npy", bestweight)
 # bestweight = np.load(CurrentPath+"bestweightSearch.npy")
-# utils.NumpyToImage(bestweight, CurrentPath+"image")
+# utils.NumpyToImage(bestweight, CurrentPath+"image",title="SearchWerightBest")
 # EvaluatorUnsuper.SetModelConvWeight(model,layer,bestweight)
 # utils.SaveModel(model,CurrentPath+"/checkpointSearch.pkl")
 
@@ -90,12 +104,12 @@ utils.NumpyToImage(weight, CurrentPath+"image",title="SearchWeight")
 
 # weight = EvaluatorUnsuper.UnsuperLearnTrainWeight(model, layer, traindata, NumTrain=5000)
 # np.save("WeightTrain.npy", weight)
-# utils.NumpyToImage(bestweight, CurrentPath+"image",title="TrainWeight")
 # weight = np.load(CurrentPath+"WeightTrain.npy")
+# utils.NumpyToImage(weight, CurrentPath+"image",title="TrainWeight")
 # bestweight, index = EvaluatorUnsuper.UnsuperLearnFindBestWeight(model, layer, weight, traindata, databatchs=64, interation=1000000)
 # np.save(CurrentPath+"bestweightTrain.npy", bestweight)
 # bestweight = np.load(CurrentPath+"bestweightTrain.npy")
-# utils.NumpyToImage(bestweight, CurrentPath+"image")
+# utils.NumpyToImage(bestweight, CurrentPath+"image",title="TrainWerightBest")
 # EvaluatorUnsuper.SetModelConvWeight(model,layer,bestweight)
 # utils.SaveModel(model,CurrentPath+"/checkpointTrain.pkl")
 

FilterEvaluator/EvaluatorUnsuper.py

@@ -66,8 +66,6 @@ def UnsuperLearnSearchWeight(model, layer, dataloader, NumSearch=10000, SaveChan
 
     for i in range(NumSearch):
         newlayer.weight.data=utils.SetDevice(torch.from_numpy(newweight[i]))
-
-
         output = model.ForwardLayer(dataset,layer-1)
         output = newlayer(output)
         output = torch.reshape(output.transpose(0,1),(newlayer.out_channels,-1))
@@ -78,11 +76,8 @@ def UnsuperLearnSearchWeight(model, layer, dataloader, NumSearch=10000, SaveChan
         dat2 = torch.mean(dat2 * dat2,dim=1)
         score = dat2.cpu().detach().numpy()
 
-
-        # score = GetScore(model, layer, newlayer, dataset)
         minactive = np.append(minactive, score)
         minweight = np.concatenate((minweight, newweight[i]))
-
         index = minactive.argsort()
         minactive = minactive[index[0:SaveChannel]]
         minweight = minweight[index[0:SaveChannel]]
@@ -93,7 +88,7 @@ def UnsuperLearnSearchWeight(model, layer, dataloader, NumSearch=10000, SaveChan
     
     tl.data=utils.SetDevice(torch.from_numpy(minweight[0:tl.out_channels]))
     interationbar.close()
-    return minweight
+    return minweight, minactive
 
 def TrainLayer(netmodel, layer, SearchLayer, DataSet, Epoch=100):
     netmodel.eval()