Update README.md

README.md

@@ -64,4 +64,40 @@ The model was trained on a proprietary dataset of text-mined inorganic synthesis
 https://www.nature.com/articles/s41597-019-0224-1
 
 **Evaluation Results:**
-(soon)
+The models were evaluated on a hold-out test set.
+
+**1. Tuned Temperature Bin Prediction Model:**
+* **Overall Test Set Accuracy:** 0.6821
+* **Overall Test Set F1 Score (Weighted):** 0.6785
+* **Per-Class Performance (Test Set):**
+    ```
+                                  precision    recall  f1-score   support
+
+        TempBin_1_(1_to_900]       0.77      0.79      0.78       954
+     TempBin_2_(900_to_1100]       0.62      0.53      0.57       743
+    TempBin_3_(1100_to_1300]       0.58      0.58      0.58       768
+    TempBin_4_(1300_to_3000]       0.72      0.80      0.76       715
+
+                    accuracy                           0.68      3180
+                   macro avg       0.67      0.68      0.67      3180
+                weighted avg       0.68      0.68      0.68      3180
+    ```
+
+**2. Tuned Atmosphere Category Prediction Model:**
+* **Overall Test Set Accuracy:** 0.7193
+* **Overall Test Set F1 Score (Weighted):** 0.7174
+* **Per-Class Performance (Test Set):**
+    ```
+                              precision    recall  f1-score   support
+
+                   Inert       0.59      0.38      0.46       139
+        Other_Atm_Target       1.00      0.44      0.62         9
+               Oxidizing       0.67      0.71      0.69      1552
+                Reducing       0.70      0.47      0.56       100
+    Unknown_Atm_Category       0.76      0.76      0.76      2098
+
+                accuracy                           0.72      3898
+               macro avg       0.74      0.55      0.62      3898
+            weighted avg       0.72      0.72      0.72      3898
+    ```
+)