Update README.md

README.md

@@ -12,12 +12,13 @@ metrics:
 ---
 # EfficientNet-B3 Pizza/Steak/Sushi Classifier
 
-I fine-tuned a pre-trained EfficientNet-B3 model to classify images into three categories: pizza, steak, and sushi[cite: 1]. 
+I fine-tuned a pre-trained EfficientNet-B3 model to classify images into three categories: pizza, steak, and sushi. 
 
 ## Model Details
 * **Architecture:** `torchvision.models.efficientnet_b3`
 * **Weights:** `EfficientNet_B3_Weights.DEFAULT`
-* **Modifications:** I froze all the base feature layers
+* **Modifications:** I froze all the base feature layers.
+
 ## Training Procedure
 I trained the model for 10 epochs using the Adam optimizer.
 
@@ -25,7 +26,7 @@ I trained the model for 10 epochs using the Adam optimizer.
 * **Learning Rate:** 0.001
 * **Loss Function:** CrossEntropyLoss
 * **Transforms:** I used the automatic transforms provided by the default EfficientNet-B3 weights. 
-* **Hardware:** Trained using `cuda` (if available) with a set manual seed of 37 for reproducibility].
+* **Hardware:** Trained using `cuda` (if available) with a set manual seed of 37 for reproducibility.
 
 ## Dataset
 I used a 20% subset of a pizza, steak, and sushi dataset. The data was split into `train` and `test` directories. 
@@ -33,35 +34,38 @@ I used a 20% subset of a pizza, steak, and sushi dataset. The data was split int
 ## Evaluation Results
 
 ### Accuracy and Loss Curves
-Over the 10 epochs, both the training and testing loss steadily decreased, with the testing loss ending below 0.40. The testing accuracy outperformed the training accuracy early on and finished highly stable above 90%.
+Over the 10 epochs, both the training and testing loss steadily decreased, with the testing loss ending below 0.30. The testing accuracy remained highly stable and finished above 95%.
 
-![Loss and Accuracy Curves](plots/EfficientNet_B3_20percent.pth_curves.png)
+![Loss and Accuracy Curves](EfficientNet_B3_20percent.pth_curves.png)
 
 ### Confusion Matrix
 The model performs exceptionally well across all three classes on the test set:
-* **Pizza:** 45 correct, 0 misclassified as steak, 1 misclassified as sushi.
-* **Steak:** 56 correct, 0 misclassified as pizza, 2 misclassified as sushi.
-* **Sushi:** 42 correct, 3 misclassified as pizza, 1 misclassified as steak.
+* **Pizza:** 42 correct, 2 misclassified as steak, 2 misclassified as sushi.
+* **Steak:** 57 correct, 0 misclassified as pizza, 1 misclassified as sushi.
+* **Sushi:** 45 correct, 0 misclassified as pizza, 1 misclassified as steak.
 
-![Confusion Matrix](plots/EfficientNet_B3_20percent.pth_confusion_matrix.png)
+![Confusion Matrix](EfficientNet_B3_20percent.pth_confusion_matrix.png)
 
 ### Most Confident Wrong Predictions
-I plotted the instances where the model was highly confident but incorrect. The model occasionally struggled with distinguishing close-up textures, such as predicting a steak dish as sushi with 0.82 confidence, or a sushi dish as pizza with 0.61 confidence.
+I plotted the instances where the model was highly confident but incorrect. The model occasionally struggled with distinguishing close-up textures and lighting, such as predicting a steak dish as sushi with 0.73 confidence, or a pizza dish as steak with 0.46 confidence.
 
-![Wrong Predictions](plots/EfficientNet_B3_20percent.pth_wrong_pred.png)
+![Wrong Predictions](EfficientNet_B3_20percent.pth_wrong_pred.png)
 
 ## How to use
 ```python
 import torch
 import torchvision
-# 1. Load the model architecture
+
+# I loaded the model architecture
 weights = torchvision.models.EfficientNet_B3_Weights.DEFAULT
 model = torchvision.models.efficientnet_b3(weights=weights)
-# 2. Modify the classifier to match the 3 classes
+
+# I modified the classifier to match the 3 classes
 model.classifier = torch.nn.Sequential(
     torch.nn.Dropout(p=0.2, inplace=True),
-    torch.nn.Linear(in_features=1280, out_features=3, bias=True),
+    torch.nn.Linear(in_features=1536, out_features=3, bias=True),
 )
-# 3. Load the weights
+
+# I loaded the weights
 model.load_state_dict(torch.load("EfficientNet_B3_20percent.pth", map_location="cpu"))
 model.eval()