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
+09_pretrained_effnetb2_feature_extractor_20_percent.pth filter=lfs diff=lfs merge=lfs -text

09_pretrained_effnetb2_feature_extractor_20_percent.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:082946cc6f15df586b385e2c361e0f7384511deddeabcea83dbd04c2c8c71414
+size 31300218

app.py

@@ -0,0 +1,64 @@
+
+### 1. Imports and class names setup
+import gradio as gr
+import torch
+import os
+
+from timeit import default_timer as timer
+from model import create_effnetb2_model
+from typing import Tuple, Dict
+
+# Setup class names
+class_names = ["pizza", "steak", "sushi"]
+
+### 2. Model and transforms preparation
+effnetb2, effnetb2_transforms = create_effnetb2_model(num_classes = len(class_names))
+
+# Load the saved weights
+effnetb2.load_state_dict(torch.load(f="09_pretrained_effnetb2_feature_extractor_20_percent.pth",
+                                    map_location=torch.device("cpu")))
+
+### 3. Predict function
+def predict(img) -> Tuple[Dict, float]:
+    # Start a timer
+    start_time = timer()
+
+    # Transform the input image for use with EffNetB2
+    img = effnetb2_transforms(img).unsqueeze(0)
+
+    # Put model into eval mode to make prediction
+    effnetb2.eval()
+    with torch.inference_mode():
+        # Pass transformed image through the model
+        pred_probs = torch.softmax(effnetb2(img), dim=1).squeeze()
+
+    # Create a prediction label and prediction probability dictionary
+    pred_labels_and_probs = {food: float(pred_probs[i]) for i, food in enumerate(class_names)}
+
+    # Calculate pred time
+    pred_time = round(timer() - start_time, 4)
+
+    # Return pred dict and pred time
+    return pred_labels_and_probs, pred_time
+
+### 4. Create the Gradio app
+title = "FoodVision Mini🍕🥩🍣"
+description = "An [EfficientNetB2 Feature Extractor](https://pytorch.org/vision/main/models/efficientnet.html#efficientnet_b2) computer vision model to classify images as pizza, steak and sushi."
+article = "Created at [09. Pytorch Model Deployment](https://www.learnpytorch.io/09_pytorch_model_deployment)"
+
+# Create example list
+example_list = [["examples/" + example] for example in os.listdir("examples")]
+
+# Create the gradio demo
+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 the demo
+demo.launch(debug=False, # Print errors locally?
+            share=True) # generate a publically available URL

model.py

@@ -0,0 +1,24 @@
+
+import torch
+import torchvision
+
+from torch import nn
+
+def create_effnetb2_model(num_classes: int = 3,
+                          seed: int = 42):
+    # 1, 2, 3 Create EffNetB2 pretrained weights, transforms and model
+    weights = torchvision.models.EfficientNet_B2_Weights.DEFAULT
+    transforms = weights.transforms()
+    model = torchvision.models.efficientnet_b2(weights=weights)
+
+    # 4. Freeze all layers in the base model
+    for param in model.parameters():
+        param.requires_grad = False
+
+    # 5. Change classifier head with random seed for reproducibility
+    torch.manual_seed(seed)
+    model.classifier = nn.Sequential(
+        nn.Dropout(p= .3, inplace=True),
+        nn.Linear(in_features=1408, out_features=3, bias=True)
+    )
+    return model, transforms

requirements.txt

@@ -0,0 +1,3 @@
+torch=2.1.0
+torchvision=0.16.0
+gradio=3.41.0