added dropout

.gitattributes

@@ -33,3 +33,5 @@ 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
+results/pcam/19_06_2025_17_32_18/model_3.pt filter=lfs diff=lfs merge=lfs -text
+results/pcam/ filter=lfs diff=lfs merge=lfs -text

README.md

@@ -43,7 +43,7 @@ This study explores the application of DenseNet architectures to the PCam datase
 
 The submission on kaggle with the model trained on this notebook is 
 
-```Public score: 0.9611```
+```Public score: 0.9626```
 
 ## ⚡ Try it now ! With gradio ⚡ 
 

app.py

@@ -11,7 +11,7 @@ from PIL import Image
 # ---------------------------------
 torch.manual_seed(42)
 torch.cuda.manual_seed_all(42)
-model = torch.load("results/pcam/17_06_2025_12_19_40/model_5.pt", map_location="cpu", weights_only=False)
+model = torch.load("results/pcam/19_06_2025_17_32_18/model_3.pt", map_location="cpu", weights_only=False)
 model.eval()
 
 # ---------------------------------
@@ -46,7 +46,7 @@ def get_sample(index: int):
     with torch.no_grad():
         output = model(image_tensor.unsqueeze(0)).squeeze()
         probability = torch.sigmoid(output)
-        predicted_label = "Tumor" if probability >= 0.45 else "No Tumor"
+        predicted_label = "Tumor" if probability >= 0.4096705 else "No Tumor"
         true_label = "Tumor" if ground_truth == 1 else "No Tumor"
         error_label = ""
         if predicted_label != true_label: