Upload 5 files

app.py

@@ -0,0 +1,90 @@
+### 1. Imports and class names setup ### 
+import gradio as gr
+import os
+import torch
+from model import TinyCNN
+from timeit import default_timer as timer
+from typing import Tuple, Dict
+import torch
+import torchvision
+from torchvision import transforms
+from torch import nn
+# Setup class names
+with open("class_names.txt", "r") as f: # reading them in from class_names.txt
+    class_names = [defects.strip() for defects in  f.readlines()]
+    
+### 2. Model and transforms preparation ###    
+
+# Create model
+TinyCNN_model = TinyCNN(input_shape=3,
+                  hidden_units=64, 
+                  output_shape=len(class_names))
+
+
+transform = transforms.Compose([
+    transforms.Resize((224, 224)),
+    transforms.ToTensor()
+])
+
+# Load saved weights
+TinyCNN_model.load_state_dict(
+    torch.load(
+        f="cnn.pth",
+        map_location=torch.device("cpu"),  # load to CPU
+    )
+)
+
+### 3. Predict function ###
+
+# Create predict function
+def predict(img) -> Tuple[Dict, float]:
+    """Transforms and performs a prediction on img and returns prediction and time taken.
+    """
+    # Start the timer
+    start_time = timer()
+    
+    # Transform the target image and add a batch dimension
+    img = transform(img).unsqueeze(dim=0)
+    
+    # Put model into evaluation mode and turn on inference mode
+    TinyCNN_model.eval()
+    with torch.inference_mode():
+        # Pass the transformed image through the model and turn the prediction logits into prediction probabilities
+        pred_probs = torch.softmax(TinyCNN_model(img), dim=1)
+    
+    # Create a prediction label and prediction probability dictionary for each prediction class (this is the required format for Gradio's output parameter)
+    pred_labels_and_probs = {class_names[i]: float(pred_probs[0][i]) for i in range(len(class_names))}
+    
+    # Calculate the prediction time
+    pred_time = round(timer() - start_time, 5)
+    
+    # Return the prediction dictionary and prediction time 
+    return pred_labels_and_probs, pred_time
+
+
+### 4. Gradio app ###
+
+# Create title, description and article strings
+title = "Expression Detection"
+description = "An app to predict emotions from the list.[Angry, Disgust, Fear, Happy, Neutral, Sad, Surprise]"
+article = "Created as a college  project."
+
+# Create examples list from "examples/" directory
+example_list = [["examples/" + example] for example in os.listdir("examples")]
+
+# Create Gradio interface 
+demo = gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil"),
+    outputs=[
+        gr.Label(num_top_classes=5, label="Predictions"),
+        gr.Image(label="Defects"),
+    ],
+    examples=example_list,
+    title=title,
+    description=description,
+    article=article,
+)
+
+# Launch the app!
+demo.launch()

class_names.txt

@@ -0,0 +1,9 @@
+Center
+Donut
+Edge-Loc
+Edge-Ring
+Loc
+Near-full
+Random
+Scratch
+none

cnn.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be41344ebccd7a18bdaa26887f89c62324ee8b6a3032744eb0023a6b215be216
+size 115453650

model.py

@@ -0,0 +1,91 @@
+import torch
+import torchvision
+
+from torch import nn
+
+class TinyCNN(nn.Module):
+
+  def __init__(self, input_shape: int, hidden_units: int, output_shape: int) -> None:
+      super().__init__()
+      self.conv_block_1 = nn.Sequential(
+        nn.Conv2d(in_channels=input_shape,
+                out_channels=hidden_units,
+                kernel_size=3, # how big is the square that's going over the image?
+                stride=1, # default
+                padding=1), # options = "valid" (no padding) or "same" (output has same shape as input) or int for specific number
+        nn.BatchNorm2d(hidden_units),
+        nn.ReLU(),
+        # nn.Conv2d(in_channels=hidden_units,
+        #         out_channels=128,
+        #         kernel_size=3,
+        #         stride=1,
+        #         padding=0),
+        # nn.BatchNorm2d(128),
+        # nn.ReLU(),
+        nn.MaxPool2d(kernel_size=2,
+                        stride=2), # default stride value is same as kernel_size
+        nn.Dropout(p=0.25)
+      )
+      self.conv_block_2 = nn.Sequential(
+          nn.Conv2d(hidden_units, 128, kernel_size=3, padding=1),
+          # nn.ReLU(),
+          # nn.Conv2d(128, 128, kernel_size=3, padding=0),
+          nn.BatchNorm2d(128),
+          nn.ReLU(),
+          nn.MaxPool2d(2),
+          nn.Dropout(p=0.25)
+      )
+
+      self.conv_block_3 = nn.Sequential(
+          nn.Conv2d(128, 512, kernel_size=3, padding=1),
+          # nn.ReLU(),
+          # nn.Conv2d(128, 512, kernel_size=3, padding=0),
+          nn.BatchNorm2d(512),
+          nn.ReLU(),
+          nn.MaxPool2d(2),
+          nn.Dropout(p=0.25)
+      )
+
+      self.conv_block_4 = nn.Sequential(
+          nn.Conv2d(512, 512, kernel_size=3, padding=1),
+        #   nn.ReLU(),
+        #   nn.Conv2d(512, 512, kernel_size=3, padding=2),
+          nn.BatchNorm2d(512),
+          nn.ReLU(),
+          nn.MaxPool2d(2),
+          nn.Dropout(p=0.25)
+      )
+
+      self.fc_1 = nn.Sequential(
+          nn.Flatten(),
+          nn.Linear(in_features=256*392, out_features = 256),
+          nn.BatchNorm1d(256),
+          nn.ReLU(),
+          nn.Dropout(p=0.25)
+      )
+
+      self.fc_2 = nn.Sequential(
+
+          # Where did this in_features shape come from?
+          # It's because each layer of our network compresses and changes the shape of our inputs data.
+          nn.Linear(in_features=256,
+                    out_features=512),
+          nn.BatchNorm1d(512),
+          nn.ReLU(),
+          nn.Dropout(p=0.25)
+      )
+
+      self.classifier = nn.Sequential(
+          nn.Linear(in_features=512,
+                  out_features=output_shape)
+      )
+
+  def forward(self, x):
+      x = self.conv_block_1(x)
+      x = self.conv_block_2(x)
+      x = self.conv_block_3(x)
+      x = self.conv_block_4(x)
+      x = self.fc_1(x)
+      x = self.fc_2(x)
+      x = self.classifier(x)
+      return x

requirements.txt

@@ -0,0 +1,3 @@
+torch==2.2.0
+torchvision==0.17.0
+gradio==4.17.0