Interesting case.

demo/guide-python/categorical.py

@@ -49,31 +49,35 @@ def main() -> None:
     # Use builtin categorical data support
     # For scikit-learn interface, the input data must be pandas DataFrame or cudf
     # DataFrame with categorical features
-    X, y = make_categorical(100, 10, 4, False)
+    X, y = make_categorical(100, 1, 2, False)
+    print(X)
     # Specify `enable_categorical` to True.
     tree_method = "approx"
-    reg = xgb.XGBRegressor(tree_method=tree_method, enable_categorical=True)
+    reg = xgb.XGBRegressor(
+        tree_method=tree_method, enable_categorical=True,
+        max_depth=1, n_estimators=1
+    )
     reg.fit(X, y, eval_set=[(X, y)])
 
     # Pass in already encoded data
-    X_enc, y_enc = make_categorical(100, 10, 4, True)
-    reg_enc = xgb.XGBRegressor(tree_method=tree_method)
+    X_enc, y_enc = make_categorical(100, 1, 2, True)
+    reg_enc = xgb.XGBRegressor(tree_method=tree_method, max_depth=1, n_estimators=1)
     reg_enc.fit(X_enc, y_enc, eval_set=[(X_enc, y_enc)])
 
-    reg_results = np.array(reg.evals_result()["validation_0"]["rmse"])
-    reg_enc_results = np.array(reg_enc.evals_result()["validation_0"]["rmse"])
+    # reg_results = np.array(reg.evals_result()["validation_0"]["rmse"])
+    # reg_enc_results = np.array(reg_enc.evals_result()["validation_0"]["rmse"])
 
-    # Check that they have same results
-    np.testing.assert_allclose(reg_results, reg_enc_results)
+    # # Check that they have same results
+    # np.testing.assert_allclose(reg_results, reg_enc_results)
 
-    # Convert to DMatrix for SHAP value
-    booster: xgb.Booster = reg.get_booster()
-    m = xgb.DMatrix(X, enable_categorical=True)  # specify categorical data support.
-    SHAP = booster.predict(m, pred_contribs=True)
-    margin = booster.predict(m, output_margin=True)
-    np.testing.assert_allclose(
-        np.sum(SHAP, axis=len(SHAP.shape) - 1), margin, rtol=1e-3
-    )
+    # # Convert to DMatrix for SHAP value
+    # booster: xgb.Booster = reg.get_booster()
+    # m = xgb.DMatrix(X, enable_categorical=True)  # specify categorical data support.
+    # SHAP = booster.predict(m, pred_contribs=True)
+    # margin = booster.predict(m, output_margin=True)
+    # np.testing.assert_allclose(
+    #     np.sum(SHAP, axis=len(SHAP.shape) - 1), margin, rtol=1e-3
+    # )
 
 
 if __name__ == "__main__":