diff --git a/README.md b/README.md
index 8dd4098..b9ed7ed 100644
--- a/README.md
+++ b/README.md
@@ -25,7 +25,8 @@ pixi add slim-trees
 Using `slim-trees` does not affect your training pipeline.
 Simply call `dump_sklearn_compressed` or `dump_lgbm_compressed` to save your model.
 
-> ⚠️ `slim-trees` does not save all the data that would be saved by `sklearn`:
+> [!WARNING]
+> `slim-trees` does not save all the data that would be saved by `sklearn`:
 > only the parameters that are relevant for inference are saved. If you want to save the full model including
 > `impurity` etc. for analytic purposes, we suggest saving both the original using `pickle.dump` for analytics
 > and the slimmed down version using `slim-trees` for production.
diff --git a/pixi.toml b/pixi.toml
index dd8c5f7..8a72341 100644
--- a/pixi.toml
+++ b/pixi.toml
@@ -1,6 +1,6 @@
 [project] # TODO: move to pyproject.toml once pixi supports it
 name = "slim-trees"
-version = "0.2.3"
+version = "0.2.4"
 description = "A python package for efficient pickling of ML models."
 authors = ["Pavel Zwerschke <pavelzw@gmail.com>"]
 channels = ["conda-forge"]
diff --git a/pyproject.toml b/pyproject.toml
index 2f9f317..f7989d8 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -5,7 +5,7 @@ build-backend = "hatchling.build"
 [project]
 name = "slim-trees"
 description = "A python package for efficient pickling of ML models."
-version = "0.2.3"
+version = "0.2.4"
 readme = "README.md"
 license = "MIT"
 requires-python = ">=3.8"
diff --git a/slim_trees/lgbm_booster.py b/slim_trees/lgbm_booster.py
index 30d3ec4..9da232c 100644
--- a/slim_trees/lgbm_booster.py
+++ b/slim_trees/lgbm_booster.py
@@ -198,8 +198,9 @@ def _compress_booster_handle(model_string: str) -> Tuple[str, List[dict], str]:
     return front_str, trees, back_str
 
 
-def _validate_tree_structure(tree: dict) -> bool:
-    return type(tree) == dict and tree.keys() == {
+def _validate_tree_structure(tree: dict):
+    assert isinstance(tree, dict)
+    if tree.keys() != {
         "num_leaves",
         "num_cat",
         "split_feature",
@@ -210,7 +211,11 @@ def _validate_tree_structure(tree: dict) -> bool:
         "leaf_value",
         "is_linear",
         "shrinkage",
-    }
+    }:
+        raise ValueError(
+            "Invalid tree structure. Do you use an unsupported LightGBM version or try to load a "
+            "model that was pickled with a different version of LightGBM?"
+        )
 
 
 def _decompress_booster_handle(compressed_state: Tuple[str, List[dict], str]) -> str:
diff --git a/slim_trees/sklearn_tree.py b/slim_trees/sklearn_tree.py
index 8bb456f..3c643aa 100644
--- a/slim_trees/sklearn_tree.py
+++ b/slim_trees/sklearn_tree.py
@@ -129,7 +129,7 @@ def _decompress_tree_state(state: dict):
     """
     assert isinstance(state, dict)
     # TODO: make prettier once python 3.8 is not supported anymore
-    assert state.keys() == {
+    if state.keys() != {
         *{
             "max_depth",
             "node_count",
@@ -141,7 +141,11 @@ def _decompress_tree_state(state: dict):
             "values",
         },
         *({"missing_go_to_left"} if sklearn_version >= Version("1.3") else set()),
-    }
+    }:
+        raise ValueError(
+            "Invalid tree structure. Do you use an unsupported scikit-learn version "
+            "or try to load a model that was pickled with a different version of scikit-learn?"
+        )
     n_nodes = state["node_count"]
 
     children_left = np.zeros(n_nodes, dtype=np.int64)