Update app.py with new model , update desc

app.py

@@ -14,9 +14,9 @@ class Flatten(nn.Module):
 class ConvNet(nn.Module):
     def __init__(self):
         super(ConvNet, self).__init__()
-        self.conv1 = nn.Conv2d(1, 32, kernel_size=5, stride=2, padding=1)
-        self.conv2 = nn.Conv2d(32, 32, kernel_size=5, stride=2, padding=1)
-        self.conv3 = nn.Conv2d(32, 10, kernel_size=5, stride=2, padding=1)
+        self.conv1 = nn.Conv2d(1, 64, kernel_size=5, stride=2, padding=1)
+        self.conv2 = nn.Conv2d(64, 64, kernel_size=5, stride=2, padding=1)
+        self.conv3 = nn.Conv2d(64, 10, kernel_size=5, stride=2, padding=1)
         self.fc = nn.Linear(10,10)
         self.dropout = nn.Dropout(0.5)
         self.avgpool = nn.AdaptiveAvgPool2d(1)
@@ -50,7 +50,7 @@ def predict(img, withGradio=False):
     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
     # Get the state_dict of conv_net
-    state_dict = torch.load('model-e2e-v6-1-13092023.pth')
+    state_dict = torch.load('this-is-mnist-model-f1c-desk-19092023.pth')
 
     # Define a new state_dict for ConvNet
     new_state_dict = OrderedDict()
@@ -86,8 +86,8 @@ def wrapper_fn(input_image):
     return predict(input_image, withGradio=True)
 
 # Define Gradio interface
-title = "MNIST - understanding the basis"
-description = "I have created and trained a custom neural network for MNIST. The score of `valid loss 0.126034` isn't the optimal one but it is reasonably acceptable. This is an exercise project where I only allowed myself a limited time to improve the result. You need to use same size image as the model 28x28, white text on black for best results."
+title = "MNIST - understanding the basics"
+description = "I have created and trained a CNN for MNIST. You can find the exercise notebook [here](https://www.kaggle.com/code/mindgspl/exercise-mnist). Note : use same size image as the model 28x28, white text on black for best results. "
 examples = ['data/0-custom-invert.jpg', 'data/0.jpg', 'data/2.jpg', 'data/3.jpg',  'data/5.jpg',  'data/9.jpg', 'data/0-custom.jpg',]
 output = gr.Textbox(label="Output prediction")
 app = gr.Interface(fn=wrapper_fn, inputs=gr.Image(), outputs=output, title=title,description=description,examples=examples)