diff --git a/docs/about.md b/docs/about.md
index 8171c62..94c370a 100644
--- a/docs/about.md
+++ b/docs/about.md
@@ -211,25 +211,25 @@ $$
 
         | Class    |    IoU | Precision | Recall | Accuracy |
         |:---------|-------:|----------:|-------:|---------:|
-        | Kelp     | 0.6593 |    0.8145 | 0.7804 |        - |
-        | Not Kelp | 0.9885 |    0.9939 | 0.9945 |        - |
-        | *Mean*   | 0.8239 |    0.9042 | 0.8875 |   0.9892 |
+        | Kelp     | 0.7026 |    0.7525 | 0.9122 |        - |
+        | Not Kelp | 0.9862 |    0.9978 | 0.9883 |        - |
+        | *Mean*   | 0.8444 |    0.8752 | 0.9503 |   0.9869 |
 
     === "Validation split"
 
         | Class    |    IoU | Precision | Recall | Accuracy |
         |:---------|-------:|----------:|-------:|---------:|
-        | Kelp     | 0.6964 |    0.7585 | 0.8917 |        - |
-        | Not Kelp | 0.9857 |    0.9969 | 0.9886 |        - |
-        | *Mean*   | 0.8410 |    0.8777 | 0.9402 |   0.9865 |
+        | Kelp     | 0.6892 |    0.8317 | 0.8031 |        - |
+        | Not Kelp | 0.9635 |    0.9808 | 0.9820 |        - |
+        | *Mean*   | 0.8263 |    0.9062 | 0.8926 |   0.9671 |
 
     === "Train split"
 
         | Class    |    IoU | Precision | Recall | Accuracy |
         |:---------|-------:|----------:|-------:|---------:|
-        | Kelp     | 0.7786 |    0.8510 | 0.8980 |        - |
-        | Not Kelp | 0.9716 |    0.9889 | 0.9822 |        - |
-        | *Mean*   | 0.8751 |    0.9200 | 0.9401 |   0.9757 |
+        | Kelp     | 0.7674 |    0.8446 | 0.8906 |        - |
+        | Not Kelp | 0.9724 |    0.9897 | 0.9824 |        - |
+        | *Mean*   | 0.8699 |    0.9172 | 0.9365 |   0.9757 |
 
 === "RGBI"
 
diff --git a/kelp_o_matic/models.py b/kelp_o_matic/models.py
index 841082f..a52d71d 100644
--- a/kelp_o_matic/models.py
+++ b/kelp_o_matic/models.py
@@ -93,7 +93,7 @@ class KelpRGBPresenceSegmentationModel(_Model):
     all_black_val = 0
 
     torchscript_path = (
-        "UNetPlusPlus_EfficientNetV2_m_kelp_presence_rgb_jit_dice=0.8677.pt"
+        "UNetPlusPlus_EfficientNetV2_m_kelp_presence_rgb_jit_dice=0.8703.pt"
     )
 
     def post_process(self, x: "torch.Tensor") -> "np.ndarray":