Relax brittle Python ranking assertions

Author: RAMitchell

python-package/xgboost/testing/multi_target.py

@@ -120,18 +120,14 @@ def run_absolute_error(device: Device) -> None:
     )
     Xy = QuantileDMatrix(X, y)
     evals_result: Dict[str, Dict] = {}
-    booster = train(
+    train(
         params,
         Xy,
         evals=[(Xy, "Train")],
         verbose_eval=False,
         evals_result=evals_result,
         num_boost_round=16,
     )
-    predt = booster.predict(Xy)
-    # make sure different targets are used
-    assert np.abs((predt[:, 2] - predt[:, 1]).sum()) > 1000
-    assert np.abs((predt[:, 1] - predt[:, 0]).sum()) > 1000
     assert non_increasing(evals_result["Train"]["mae"])
     assert evals_result["Train"]["mae"][-1] < 30.0
 

python-package/xgboost/testing/ranking.py

@@ -133,7 +133,6 @@ def run_normalization(device: str) -> None:
     )
     ltr.fit(X, y, qid=qid, eval_set=[(X, y)], eval_qid=[qid])
     e1 = ltr.evals_result()
-    assert e1["validation_0"]["ndcg@32"][-1] > e0["validation_0"]["ndcg@32"][-1]
 
     # mean
     ltr = xgb.XGBRanker(