Fix model loading for later versions of XGBoost.

The old method doesn't work for XGBoost versions >= 1.3 https://stackoverflow.com/a/70098541/6179774

Author: christabella

doc_source/define-pipeline.md

@@ -417,7 +417,8 @@ This section shows how to create a processing step to evaluate the accuracy of
        with tarfile.open(model_path) as tar:
            tar.extractall(path=".")
        
-       model = pickle.load(open("xgboost-model", "rb"))
+       model = xgboost.Booster()
+       model.load_model("xgboost-model")
    
        test_path = "/opt/ml/processing/test/test.csv"
        df = pd.read_csv(test_path, header=None)
@@ -687,4 +688,4 @@ A step can only appear once in either the pipeline's step list or the if/else st
 
  This pipeline definition is ready to submit to SageMaker\. In the next tutorial, you submit this pipeline to SageMaker and start an execution\. 
 
- **Next step:** [Run a pipeline](run-pipeline.md) 
+ **Next step:** [Run a pipeline](run-pipeline.md)