Upload README.md with huggingface_hub

README.md

@@ -26,27 +26,31 @@ This model is a **Siamese UNet** designed for binary change detection using fuse
 The architecture uses dual weight-shared **ResNet-34** encoders to extract multi-modal features from pre-event RGB (EO) and post-event grayscale (SAR) images. Feature differences are fused via skip connections into a UNet decoder.
 
 - **Encoder:** ResNet-34 (Pre-trained on ImageNet)
-- **Weights Format:** Safetensors (Modern, fast, and secure)
-- **Loss Function:** Combined Focal Loss + Dice Loss (Optimized for 58:1 class imbalance)
+- **Weights Format:** Safetensors
+- **Loss Function:** Combined Focal Loss + Dice Loss
 - **Input Resolution:** 256x256
 
-## Intended Uses
-This model is intended for rapid disaster damage assessment where pre-event optical data and post-event SAR data are available.
-
-## Training Results (Validation Set)
-- **IoU:** ~27.73%
-- **F1 Score:** ~43.42%
-- **Precision:** ~35.79%
-- **Recall:** ~55.18%
+## Training Results (Final Run)
+- **Best Validation IoU:** 24.74%
+- **Precision:** 26.45%
+- **Recall:** 79.29%
+- **F1 Score:** 39.67%
 
 ## How to Use
-The model expects a 3-channel EO image (Pre-event) and a 1-channel SAR image (Post-event) as inputs.
+To use this model, ensure you have the `src/` folder from the [GitHub repository](https://github.com/rishii100/Binary-Change-Detection) in your local directory.
 
 ```python
+import torch
 from safetensors.torch import load_file
-# from src.model import SiameseUNet
+from src.model import SiameseUNet
+
+# 1. Initialize architecture
+model = SiameseUNet(pretrained=False)
+
+# 2. Load professional weights
+weights = load_file("model.safetensors")
+model.load_state_dict(weights)
 
-# model = SiameseUNet()
-# weights = load_file("model.safetensors")
-# model.load_state_dict(weights)
+# 3. Ready for inference
+model.eval()
 ```