39390cc2ee
- A `DeviceOrd` struct is implemented to indicate the device. It will eventually replace the `gpu_id` parameter. - The `predictor` parameter is removed. - Fallback to `DMatrix` when `inplace_predict` is not available. - The heuristic for choosing a predictor is only used during training.
70 lines
2.4 KiB
Python
70 lines
2.4 KiB
Python
"""Loading a pickled model generated by test_pickling.py, only used by
|
|
`test_gpu_with_dask.py`"""
|
|
import json
|
|
import os
|
|
|
|
import numpy as np
|
|
import pytest
|
|
from test_gpu_pickling import build_dataset, load_pickle, model_path
|
|
|
|
import xgboost as xgb
|
|
from xgboost import testing as tm
|
|
|
|
|
|
class TestLoadPickle:
|
|
def test_load_pkl(self) -> None:
|
|
"""Test whether prediction is correct."""
|
|
assert os.environ["CUDA_VISIBLE_DEVICES"] == "-1"
|
|
bst = load_pickle(model_path)
|
|
x, y = build_dataset()
|
|
if isinstance(bst, xgb.Booster):
|
|
test_x = xgb.DMatrix(x)
|
|
res = bst.predict(test_x)
|
|
else:
|
|
res = bst.predict(x)
|
|
assert len(res) == 10
|
|
bst.set_params(n_jobs=1) # triggers a re-configuration
|
|
res = bst.predict(x)
|
|
|
|
assert len(res) == 10
|
|
|
|
def test_context_is_removed(self) -> None:
|
|
"""Under invalid CUDA_VISIBLE_DEVICES, context should reset"""
|
|
assert os.environ["CUDA_VISIBLE_DEVICES"] == "-1"
|
|
bst = load_pickle(model_path)
|
|
config = bst.save_config()
|
|
config = json.loads(config)
|
|
assert config["learner"]["generic_param"]["gpu_id"] == "-1"
|
|
|
|
def test_context_is_preserved(self) -> None:
|
|
"""Test the device context is preserved after pickling."""
|
|
assert "CUDA_VISIBLE_DEVICES" not in os.environ.keys()
|
|
bst = load_pickle(model_path)
|
|
config = bst.save_config()
|
|
config = json.loads(config)
|
|
assert config["learner"]["generic_param"]["gpu_id"] == "0"
|
|
|
|
def test_wrap_gpu_id(self) -> None:
|
|
assert os.environ["CUDA_VISIBLE_DEVICES"] == "0"
|
|
bst = load_pickle(model_path)
|
|
config = bst.save_config()
|
|
config = json.loads(config)
|
|
assert config["learner"]["generic_param"]["gpu_id"] == "0"
|
|
|
|
x, y = build_dataset()
|
|
test_x = xgb.DMatrix(x)
|
|
res = bst.predict(test_x)
|
|
assert len(res) == 10
|
|
|
|
def test_training_on_cpu_only_env(self) -> None:
|
|
assert os.environ["CUDA_VISIBLE_DEVICES"] == "-1"
|
|
rng = np.random.RandomState(1994)
|
|
X = rng.randn(10, 10)
|
|
y = rng.randn(10)
|
|
with tm.captured_output() as (out, err):
|
|
# Test no thrust exception is thrown
|
|
with pytest.raises(xgb.core.XGBoostError):
|
|
xgb.train({"tree_method": "gpu_hist"}, xgb.DMatrix(X, y))
|
|
|
|
assert out.getvalue().find("No visible GPU is found") != -1
|