deploy to huggingface

.gitignore

@@ -0,0 +1,2 @@
+data/*
+model/*

NeuralNet.py

@@ -0,0 +1,14 @@
+import torch
+import torch.nn as nn
+class NeuralNet(nn.Module):
+    def __init__(self, input_size, hidden_size, num_classes):
+        super(NeuralNet, self).__init__()
+        self.l1 = nn.Linear(input_size, hidden_size)
+        self.relu = nn.ReLU()
+        self.l2 = nn.Linear(hidden_size, num_classes)
+    
+    def forward(self, x):
+        out = self.l1(x)
+        out = self.relu(out)
+        out = self.l2(out)
+        return out

app.py

@@ -0,0 +1,55 @@
+import torch
+import torch.nn as nn
+import torchvision.transforms as transforms
+from PIL import Image
+import gradio as gr
+from NeuralNet import NeuralNet
+
+# Device Config
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# Model Configurations
+input_size = 784  # 28x28
+hidden_size = 100
+num_classes = 10
+
+# Load the trained model (Assuming you have a trained model saved as 'model.pth')
+model = NeuralNet(input_size, hidden_size, num_classes)
+model.load_state_dict(torch.load('model/model.pt', map_location=device))
+model.to(device)
+model.eval()
+
+# Define the transform
+transform = transforms.Compose([
+    transforms.Grayscale(num_output_channels=1),
+    transforms.Resize((28, 28)),
+    transforms.ToTensor(),
+    transforms.Normalize((0.1307,), (0.3081,))
+])
+
+# Gradio function to process the image and make predictions
+def predict(image):
+    # Load the image
+    image = Image.fromarray(image)
+
+    # Preprocess the image
+    image = transform(image).unsqueeze(0).to(device)
+    image = image.view(-1, 28*28)  # Flatten the image
+
+    # Make prediction
+    with torch.no_grad():
+        outputs = model(image)
+        _, predicted = torch.max(outputs.data, 1)
+        return int(predicted.item())
+
+# Create a Gradio interface
+interface = gr.Interface(fn=predict, 
+                         inputs=gr.Image(), 
+                         outputs="label", 
+                         live=False, 
+                         title="Digit Recognizer using Feed-Forward Nueral Network",
+                         description="Upload a digit image to recognize it")
+
+# Launch the interface
+if __name__ == "__main__":
+    interface.launch()

train.py

@@ -0,0 +1,80 @@
+import torch
+import torch.nn as nn
+import torchvision
+import torchvision.transforms as transforms
+import matplotlib.pyplot as plt
+from NeuralNet import NeuralNet
+
+# Device Config
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# hyper parameters
+input_size = 784 # 28*28
+hidden_size = 100
+num_classes = 10
+num_epochs = 20
+batch_size = 500
+learning_rate = 0.001
+
+# MNIST
+training_dataset = torchvision.datasets.MNIST(root='./data', train=True,
+                                              transform=transforms.ToTensor(), download=True)
+test_dataset = torchvision.datasets.MNIST(root='./data', train=False,
+                                          transform=transforms.ToTensor())
+train_loader = torch.utils.data.DataLoader(dataset=training_dataset, batch_size=batch_size, shuffle=True)
+test_loader = torch.utils.data.DataLoader(dataset=test_dataset, batch_size=batch_size, shuffle=False)
+
+example = iter(train_loader)
+samples, labels = next(example)
+print(samples.shape, labels.shape)
+
+# for i in range(6):
+#     plt.subplot(2, 3, i+1)
+#     plt.imshow(samples[i][0], cmap='gray')
+# plt.show()
+
+model = NeuralNet(input_size, hidden_size, num_classes)
+
+#loss and optimizer
+criterion = nn.CrossEntropyLoss()
+optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)
+
+#training loop
+n_total_steps = len(train_loader)
+for epoch in range(num_epochs):
+    for i, (images, labels) in enumerate(train_loader):
+        # 100, 1, 28, 28
+        # n, c, h, w
+        images = images.reshape(-1, 28*28).to(device)
+        labels = labels.to(device)
+
+        #forward
+        outputs = model(images)
+        loss = criterion(outputs, labels)
+
+        #backward
+        optimizer.zero_grad()
+        loss.backward()
+        optimizer.step()
+
+        if (i+1) % 100 == 0:
+            print(f'epoch {epoch+1}/{num_epochs}, step {i+1}/{n_total_steps}, loss = {loss.item():.4f}')
+
+# test
+with torch.no_grad():
+    n_correct = 0
+    n_samples = 0
+    for images , labels in test_loader:
+        images = images.reshape(-1, 28*28).to(device)
+        labels = labels.to(device)
+
+        outputs = model(images)
+
+        # value, index
+        _, predictions = torch.max(outputs, 1)
+        n_samples += labels.shape[0]
+        n_correct += (predictions == labels).sum().item()
+
+    acc = 100.0 * n_correct / n_samples
+    print(f'accuracy = {acc}')
+torch.save(model.state_dict(), 'model/model.pt')