Upload README.md with huggingface_hub

README.md

@@ -7,14 +7,36 @@ tags:
 - thyroid
 - ultrasound
 - convnext
+- explainable-ai
 metrics:
-- auc: 0.8843
-- sensitivity: 0.7615
+- auc: 0.9362
+- accuracy: 0.8681
 ---
-# Thyroid Nodule Malignancy Detector
-This model uses a Dual-Stream Spatial-Frequency Fusion architecture (ConvNeXt-Tiny + FFT).
+# Thyroid Nodule Malignancy Detector (5-Fold Validated)
 
-### Performance Metrics (Fold 0):
-- AUC: 0.8843
-- Sensitivity: 0.7615
-- Specificity: 0.8421
+This model uses a **Dual-Stream Spatial-Frequency Fusion architecture** (ConvNeXt-Tiny + FFT Magnitude Spectrum) to classify thyroid nodules in ultrasound images.
+
+## 📊 5-Fold Cross-Validation Performance
+The model was evaluated using a stratified 5-fold cross-validation on the consolidated 7,058 image dataset.
+
+|   Fold |   Accuracy |      AUC |   Sensitivity |   Specificity |
+|-------:|-----------:|---------:|--------------:|--------------:|
+|      1 |   0.884561 | 0.94365  |      0.912134 |      0.826754 |
+|      2 |   0.86898  | 0.936338 |      0.881799 |      0.842105 |
+|      3 |   0.875354 | 0.944329 |      0.912134 |      0.798246 |
+|      4 |   0.860383 | 0.932138 |      0.876569 |      0.826374 |
+|      5 |   0.851169 | 0.924778 |      0.872385 |      0.806593 |
+
+### Summary Statistics:
+- **Mean AUC:** 0.9362 ± 0.0082
+- **Mean Sensitivity:** 0.8910 ± 0.0196
+- **Mean Specificity:** 0.8200 ± 0.0175
+
+## 🚀 Clinical Application
+This model is designed for clinical decision support. The weights hosted here (pytorch_model.bin) correspond to **Fold 3**, which achieved the highest individual AUC of 0.9443.
+
+## 🛠 Methodology
+- **Backbone:** ConvNeXt-Tiny (Spatial Stream)
+- **Texture Analysis:** FFT Magnitude Spectrum (Frequency Stream)
+- **Preprocessing:** CLAHE (Contrast Limited Adaptive Histogram Equalization)
+- **Loss Function:** Focal Loss (α=1, γ=2) to handle class imbalance.