task3-4

Author: abhijeet1999

Tasks/daily tasks/Abhijeet/image/index.jpeg

@@ -0,0 +1,195 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torchvision
+import torchvision.transforms as transforms
+import torch.optim as optim
+
+transform = transforms.Compose(
+    [
+        transforms.ToTensor(),
+        transforms.Normalize(
+            (0.5, 0.5, 0.5),
+            (0.5, 0.5, 0.5)
+        )
+    ]
+)
+
+trainset = torchvision.datasets.CIFAR10(
+    root='./data',
+    train=True,
+    download=False,
+    transform=transform
+)
+
+testset = torchvision.datasets.CIFAR10(
+    root='./data',
+    train=False,
+    download=False,
+    transform=transform
+)
+
+trainloader = torch.utils.data.DataLoader(
+    trainset,
+    batch_size=4,
+    shuffle=True,
+    num_workers=2
+)
+
+testloader = torch.utils.data.DataLoader(
+    testset,
+    batch_size=4,
+    shuffle=False,
+    num_workers=2
+)
+
+classes = (
+    'plane', 'car', 'bird', 'cat',
+           'deer', 'dog', 'frog', 'horse', 'ship', 'truck'
+)
+
+class Net(nn.Module):
+    def __init__(self):
+        super(Net, self).__init__()
+        self.conv1 = nn.Conv2d(3, 6, 5)
+        self.pool = nn.MaxPool2d(2, 2)
+        self.conv2 = nn.Conv2d(6, 16, 5)
+        self.fc1 = nn.Linear(16 * 5 * 5, 120)
+        self.fc2 = nn.Linear(120, 84)
+        self.fc3 = nn.Linear(84, 10)
+
+    def forward(self, x):
+        x = self.pool(F.relu(self.conv1(x)))
+        x = self.pool(F.relu(self.conv2(x)))
+        x = x.view(-1, 16 * 5 * 5)
+        x = F.relu(self.fc1(x))
+        x = F.relu(self.fc2(x))
+        x = self.fc3(x)
+        return x
+
+
+net = Net()
+
+loss_function = nn.CrossEntropyLoss()
+optimizer = optim.SGD(
+    net.parameters(),
+    lr=0.001
+)
+
+for epoch in range(2):
+    running_loss = 0.0
+    for i, data in enumerate(trainloader, 0):
+        # data = (inputs, labels)
+        inputs, labels = data
+        optimizer.zero_grad()
+
+        outputs = net(inputs)
+        loss = loss_function(outputs, labels)
+        loss.backward()
+        optimizer.step()
+
+        running_loss = running_loss + loss.item()
+        if i % 2000 == 1999:
+            print(
+                '[%d, %5d] loss: %.3f' %
+                (epoch + 1, i+1, running_loss/2000)
+            )
+            running_loss = 0.0
+print("vola")
+
+'''
+OUTPUT:
+ number of output channels=16, kernel size=5, learning rate=0.001 epoch=2
+final loss ===========>[2, 12000] loss: 1.811
+Test Accuracy of plane: 50% (502/1000)
+Test Accuracy of   car: 49% (495/1000)
+Test Accuracy of  bird: 15% (157/1000)
+Test Accuracy of   cat: 15% (151/1000)
+Test Accuracy of  deer: 36% (362/1000)
+Test Accuracy of   dog: 22% (222/1000)
+Test Accuracy of  frog: 49% (495/1000)
+Test Accuracy of horse: 34% (348/1000)
+Test Accuracy of  ship: 52% (520/1000)
+Test Accuracy of truck: 42% (426/1000)
+
+Test Accuracy (Overall): 36% (3678/10000)
+
+ number of output channels=16, kernel size=5, learning rate=0.01 epoch=2
+final loss ===========>[2, 12000] loss: 1.255
+
+Test Accuracy of plane: 60% (609/1000)
+Test Accuracy of   car: 71% (719/1000)
+Test Accuracy of  bird: 45% (453/1000)
+Test Accuracy of   cat: 44% (445/1000)
+Test Accuracy of  deer: 46% (469/1000)
+Test Accuracy of   dog: 42% (423/1000)
+Test Accuracy of  frog: 68% (681/1000)
+Test Accuracy of horse: 65% (659/1000)
+Test Accuracy of  ship: 66% (660/1000)
+Test Accuracy of truck: 34% (344/1000)
+
+Test Accuracy (Overall): 54% (5462/10000)
+
+number of output channels=16, kernel size=3 learning rate=0.01 epoch=2
+final loss ===========>[2, 12000] loss: 1.270
+Test Accuracy of plane: 63% (632/1000)
+Test Accuracy of   car: 74% (747/1000)
+Test Accuracy of  bird: 41% (419/1000)
+Test Accuracy of   cat: 36% (366/1000)
+Test Accuracy of  deer: 61% (610/1000)
+Test Accuracy of   dog: 30% (306/1000)
+Test Accuracy of  frog: 74% (740/1000)
+Test Accuracy of horse: 62% (624/1000)
+Test Accuracy of  ship: 62% (621/1000)
+Test Accuracy of truck: 49% (491/1000)
+
+Test Accuracy (Overall): 55% (5556/10000)
+
+number of output channels=32, kernel size=3 learning rate=0.01 epoch=2
+final loss ===========>[2, 12000] loss: 1.170
+Test Accuracy of plane: 64% (644/1000)
+Test Accuracy of   car: 67% (673/1000)
+Test Accuracy of  bird: 37% (376/1000)
+Test Accuracy of   cat: 59% (593/1000)
+Test Accuracy of  deer: 60% (603/1000)
+Test Accuracy of   dog: 38% (381/1000)
+Test Accuracy of  frog: 70% (706/1000)
+Test Accuracy of horse: 64% (646/1000)
+Test Accuracy of  ship: 70% (707/1000)
+Test Accuracy of truck: 72% (728/1000)
+
+Test Accuracy (Overall): 60% (6057/10000)
+
+number of output channels=32, kernel size=3 learning rate=0.01 epoch=4
+final loss ===========>[[4, 12000] loss: 0.966
+
+number of output channels=32, kernel size=3 learning rate=0.01 epoch=4 batch 5
+final loss ===========>[[6, 10000] loss: 0.799
+Test Accuracy of plane: 60% (606/1000)
+Test Accuracy of   car: 83% (831/1000)
+Test Accuracy of  bird: 49% (498/1000)
+Test Accuracy of   cat: 38% (385/1000)
+Test Accuracy of  deer: 71% (712/1000)
+Test Accuracy of   dog: 69% (690/1000)
+Test Accuracy of  frog: 73% (733/1000)
+Test Accuracy of horse: 68% (682/1000)
+Test Accuracy of  ship: 80% (808/1000)
+Test Accuracy of truck: 65% (651/1000)
+
+Test Accuracy (Overall): 65% (6596/10000)
+
+number of output channels=32, kernel size=3 learning rate=0.01 epoch=4 batch 5 with padding 
+final loss ===========>[6, 10000] loss: 0.718
+Test Accuracy of plane: 68% (689/1000)
+Test Accuracy of   car: 64% (645/1000)
+Test Accuracy of  bird: 57% (573/1000)
+Test Accuracy of   cat: 48% (481/1000)
+Test Accuracy of  deer: 58% (588/1000)
+Test Accuracy of   dog: 63% (633/1000)
+Test Accuracy of  frog: 83% (838/1000)
+Test Accuracy of horse: 65% (653/1000)
+Test Accuracy of  ship: 82% (829/1000)
+Test Accuracy of truck: 75% (755/1000)
+
+Test Accuracy (Overall): 66% (6684/10000)
+'''

Tasks/daily tasks/Abhijeet/task3.py

@@ -0,0 +1,32 @@
+
+
+from torchvision import  transforms
+from PIL import Image
+
+import torchvision.transforms.functional as F
+import torch
+
+
+
+transform = transforms.Compose([
+transforms.Resize(255),
+transforms.CenterCrop(224),
+transforms.ColorJitter(brightness=1, contrast=1, saturation=0, hue=0),
+transforms.RandomVerticalFlip(),
+   transforms.ToTensor(),
+      transforms.Normalize(
+            (0.5, 0.5, 0.5),
+            (0.5, 0.5, 0.5)
+        ),
+      
+
+])
+
+
+img=Image.open('image/index.jpeg')
+
+img = transform(img)
+
+
+a = F.to_pil_image(img)
+a.show()