Update README.md

README.md

@@ -14,13 +14,42 @@ tags:
 - geotechnics
 ---
 
-Multi-label classification model for visual analysis of soil core samples, trained / validated / tested on `grano1/core_sample_image_data` dataset. The model aims to predict one of the following four classes labels from an image:
+# Core Sample Main Fraction Classifier
+
+This multi-label classification model was fine-tuned for automated visual analysis of soil core samples, and trained / validated / tested on the `grano1/core_sample_image_data` dataset available via the HuggingFace eco-system. The model aims to predict one of the following four class labels from cropped core sample images (300x300 pixels):
 - Clay (T)
 - Silt (U)
 - Sand (S)
 - Gravel (G)
 
-TODO:
-- metrics
-- architecture and training details
-- limitations
+## Metrics
+
+The following table summarizes selected metrics quantifying the model performance on the test set:
+
+| Metric                             | Value  |
+|------------------------------------|--------|
+| Categorical Cross-Entropy Loss     | 0.0998 |
+| Accuracy                           | 0.9606 |
+| F1-score (macro, aggregated)       | 0.8195 |
+| F1-score (weighted, aggregated)    | 0.9597 |
+
+## Hyperparameters
+
+The following table summarizes selected hyperparameters used for training the model:
+
+| Metric                    | Value    |
+|---------------------------|----------|
+| Batch size                | 32       |
+| Optimizer                 | AdamW    |
+| Warm-up ratio             | 0.1      |
+| Metric for best model     | Cat. CE  |
+| Early stopping patience   | 3        |
+
+## Usage
+
+```python
+
+## ✨ Citation
+If you use this model, please cite:
+
+@thesis{Granitzer2025, author = {Andreas-Nizar Granitzer}, title = {Data-centric core sample classification and constitutive model selection in geotechnics}, school = {University of Applied Sciences Kufstein}, year = {2025}, type = {Master's thesis} }