Initial commit

.gitignore

@@ -0,0 +1,3 @@
+
+.venv/
+__pycache__/

.vscode/settings.json

@@ -0,0 +1,6 @@
+{
+	"[python]": {
+		"editor.defaultFormatter": "ms-python.autopep8"
+	},
+	"python.formatting.provider": "none"
+}

README.md

@@ -1,6 +1,6 @@
 ---
-title: Cnn Pytorch Mnist
-emoji: 🏢
+title: CNN MNIST Classifier
+emoji: 🔢
 colorFrom: blue
 colorTo: red
 sdk: gradio
@@ -10,4 +10,4 @@ pinned: false
 license: mit
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+The model used is a Convolutional Neural Network made with PyTorch.

app.py

@@ -0,0 +1,42 @@
+import gradio as gr
+import pickle
+import torch
+import numpy as np
+
+device = 'cuda' if torch.cuda.is_available() else 'cpu'
+
+with open('cnn_model_9961_sm.bin', 'rb') as f:
+    nn = pickle.load(f)
+
+nn.to(device)
+
+
+def predict(input):
+    if input is None:
+        return 'None'
+
+    x = np.array([[input]])
+    x = torch.tensor(x).to(device)
+    p = nn(x)
+    p = p[0].cpu().detach().numpy()
+
+    return dict(enumerate(p.tolist()))
+
+
+demo = gr.Interface(
+    fn=predict,
+    inputs=[
+        gr.Sketchpad(
+            shape=(28, 28),
+            brush_radius=1.2,
+        )
+    ],
+    outputs=[
+        gr.Label(
+            num_top_classes=3,
+            scale=2,
+        )
+    ],
+    live=True,
+    allow_flagging='never',
+).launch()

networktorch.py

@@ -0,0 +1,50 @@
+from torch import nn
+
+
+class NeuralNetworkTorch(nn.Module):
+    def __init__(self):
+        super().__init__()
+
+        self.stack = nn.Sequential(
+            nn.Linear(784, 64),
+            nn.Sigmoid(),
+
+            nn.Linear(64, 10),
+            nn.Sigmoid()
+        )
+
+    def forward(self, x):
+        return self.stack(x)
+
+
+class ConvNeuralNetworkTorch(nn.Module):
+    def __init__(self):
+        super().__init__()
+
+        self.conv = nn.Sequential(
+            nn.Conv2d(1, 16, kernel_size=3, stride=1, padding=1),
+            nn.ReLU(),
+
+            nn.MaxPool2d(kernel_size=2, stride=2),
+
+            nn.Conv2d(16, 16, kernel_size=3, stride=1, padding=1),
+            nn.ReLU(),
+
+            # nn.MaxPool2d(kernel_size=2, stride=2),
+        )
+
+        self.fc = nn.Sequential(
+            nn.Linear(16 * 14 * 14, 10),
+            nn.Sigmoid(),
+        )
+
+    def forward(self, x):
+        # we do some reshaping here simply to avoid making changes to the caller
+        # so it continues to work with the fully conected network above
+        x = x.reshape(-1, 1, 28, 28) / 255
+
+        conv_output = self.conv(x)
+        flat = conv_output.reshape(len(x), -1)
+        final_output = self.fc(flat)
+
+        return final_output

requirements.txt

@@ -0,0 +1,4 @@
+numpy==1.25.2
+torch==2.0.1
+torchvision==0.15.2
+gradio==3.47.1