Improve examples (#180)

Author: Yard1

xgboost_ray/examples/higgs.py

@@ -53,7 +53,7 @@ def main():
         config,
         dtrain,
         evals_result=evals_result,
-        ray_params=RayParams(max_actor_restarts=1),
+        ray_params=RayParams(max_actor_restarts=1, num_actors=1),
         num_boost_round=100,
         evals=[(dtrain, "train")])
     taken = time.time() - start

xgboost_ray/examples/higgs_parquet.py

@@ -5,8 +5,8 @@
 import pyarrow as pa
 import pyarrow.parquet as pq
 
-from .higgs import download_higgs
-from xgboost_ray import train, RayDMatrix
+from higgs import download_higgs
+from xgboost_ray import train, RayDMatrix, RayParams
 
 FILENAME_CSV = "HIGGS.csv.gz"
 FILENAME_PARQUET = "HIGGS.parquet"
@@ -81,7 +81,7 @@ def main():
         config,
         dtrain,
         evals_result=evals_result,
-        max_actor_restarts=1,
+        ray_params=RayParams(max_actor_restarts=1, num_actors=1),
         num_boost_round=100,
         evals=[(dtrain, "train")])
     taken = time.time() - start

xgboost_ray/examples/simple_ray_dataset.py

@@ -25,15 +25,12 @@ def main(cpus_per_actor, num_actors):
     # and current Ray master API
     if hasattr(ray.data, "from_pandas_refs"):
         # Generate Ray dataset from 4 partitions
-        ray_ds = ray.data.from_pandas(np.split(data, 4))
+        ray_ds = ray.data.from_pandas(data).repartition(num_actors)
     else:
         # Split into 4 partitions
-        partitions = [ray.put(part) for part in np.split(data, 4)]
+        partitions = [ray.put(part) for part in np.split(data, num_actors)]
         ray_ds = ray.data.from_pandas(partitions)
 
-    # Generate Ray dataset from 4 partitions
-    ray_ds = ray.data.from_pandas(np.split(data, 4))
-
     train_set = RayDMatrix(ray_ds, "label")
 
     evals_result = {}