initial commit

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+finetuned_effnetb2_20percent.pth filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,68 @@
+"""The main parts are:
+1. Imports and class names setup
+2. Model and transforms preparation
+3. Write a predict function for gradio to use
+4. Write the Gradio app and the launch command
+"""
+import os
+from typing import Tuple, Dict
+import PIL
+import torch
+import torchvision
+import gradio as gr
+from timeit import default_timer
+from model import create_effnetb2_model
+
+class_names = ['pizza', 'steak', 'sushi'] #hardcoded as a list
+model, transforms = create_effnetb2_model(num_classes = len(class_names))
+
+# Load saved weights into the model, and load the model onto the CPU
+model.load_state_dict(torch.load(f = "finetuned_effnetb2_20percent.pth"),
+                      map_location = torch.device('cpu'))
+
+# Write function to run inference on gradio
+def predict(img: PIL.Image, 
+            model: nn.Module = model,
+            transforms: torchvision.transforms = transforms,
+            class_names: List[str] = class_names) -> Tuple[Dict, float]:
+    """Function to predict image class on gradio
+
+    Args:
+        img (np.array): Image as a numpy array
+        model (nn.Module, optional): Model. Defaults to vit.
+        class_names (List[str], optional): List of class anmes. Defaults to class_names.
+
+    Returns:
+        Tuple[Dict, float]: Tuplefor further processing on gradio
+    """
+    start_time = timer()
+    img = transforms(img).unsqueeze(0) #add batch dimension
+    model.eval()
+    with torch.inference_mode():
+        pred_probs = torch.softmax(model(img), dim = 1)
+    pred_labels_and_probs = {class_names[i]: float(pred_probs[0][i]) for i in range(len(class_names))}
+    end_time = timer()
+    pred_time = round(end_time - start_time, 4)
+    return pred_labels_and_probs, pred_time
+
+# Create example_list
+example_list = [["examples/" + example] for example in os.listdir("examples")]
+
+# Create Gradio App
+title = 'FoodVision Mini 🍕🥩🍣'
+description = "Using a [Vision Transformer](https://arxiv.org/abs/2010.11929) for Image Classification"
+article = "Created by [Titus Lim](https://github.com/tituslhy)"
+
+demo = gr.Interface(fn = predict,
+                    inputs = gr.Image(type = "pil"),
+                    outputs = [gr.Label(num_top_classes = 3, label = "Predictions"),
+                               gr.Number(label = "Prediction time (s)")],
+                    examples = example_list,
+                    title = title,
+                    description = description,
+                    article = article)
+
+# Launch demo
+demo.launch(debug = False, #prints errors locally
+            share = True # Generate a publicly shareable URL
+            )

finetuned_effnetb2_20percent.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b208111513233556e3b64bd5bac9ad2cdbd370613b66a973a50ef890ab87f3aa
+size 31277471

model.py

@@ -0,0 +1,18 @@
+
+import torch
+import torchvision
+from torch import nn
+
+def create_effnetb2_model(num_classes: int = 3,
+                          seed: int = 42):
+    effnetb2_weights = torchvision.models.EfficientNet_B2_Weights.DEFAULT
+    effnetb2_transforms = effnetb2_weights.transforms()
+    effnetb2 = torchvision.models.efficientnet_b2(weights=effnetb2_weights)
+    for param in effnetb2.parameters():
+        param.requires_grad = False
+    torch.manual_seed(seed)
+    effnetb2.classifier = nn.Sequential(
+        nn.Dropout(p = 0.3, inplace = True),
+        nn.Linear(in_features = 1408, out_features = num_classes)
+    )
+    return effnetb2, effnetb2_transforms