[REVIEW] Bumping xgboost version to match cuml version

[mdemoret-nv]

Upgrading the xgboost version used in the build script to match the cuML version (since its a custom build of xgboost).

This follows a similar PR for integration: rapidsai/integration#166

However, unlike the integration change, we will match the xgboost version to the cuml version to avoid hitting this issue in the future.

[GPUtester]

Please update the changelog in order to start CI tests.

View the gpuCI docs here.

[codecov-io]

Codecov Report

Merging #3062 into branch-0.17 will increase coverage by 0.84%.
The diff coverage is 89.28%.

@@               Coverage Diff               @@
##           branch-0.17    #3062      +/-   ##
===============================================
+ Coverage        58.39%   59.23%   +0.84%     
===============================================
  Files              143      142       -1     
  Lines             8897     8966      +69     
===============================================
+ Hits              5195     5311     +116     
+ Misses            3702     3655      -47     

Impacted Files	Coverage Δ
python/cuml/common/import_utils.py	61.95% <0.00%> (-2.42%)	⬇️
python/cuml/dask/solvers/cd.py	54.54% <ø> (+2.16%)	⬆️
python/cuml/thirdparty_adapters/adapters.py	88.44% <95.74%> (+6.20%)	⬆️
...l/_thirdparty/sklearn/preprocessing/_imputation.py	62.50% <100.00%> (-4.05%)	⬇️
python/cuml/thirdparty_adapters/__init__.py	100.00% <100.00%> (ø)
python/cuml/pytest_benchmarks/test_bench.py	100.00% <0.00%> (ø)
python/cuml/common/array.py	97.72% <0.00%> (+0.01%)	⬆️
python/cuml/metrics/_ranking.py	98.61% <0.00%> (+0.03%)	⬆️
python/cuml/preprocessing/encoders.py	95.18% <0.00%> (+0.07%)	⬆️
...on/cuml/_thirdparty/sklearn/preprocessing/_data.py	63.45% <0.00%> (+0.12%)	⬆️
... and 35 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ae223f1...1664d98. Read the comment docs.

[JohnZed]

I'm seeing

PASSED
cuml/test/dask/test_label_encoder.py::test_inverse_transform[orig_label3-ord_label3-expected_reverted3-bad_ord_label3-True] distributed.worker - WARNING -  Compute Failed
Function:  _inverse_transform_func
args:      (LabelEncoder(handle_unknown='error'), 0    0
1    1
2    2
3    3
4    4
dtype: int64)
kwargs:    {}
Exception: ValueError('y contains previously unseen label 4')

PASSED
cuml/test/dask/test_label_encoder.py::test_unfitted_inverse_transform PASSED
cuml/test/dask/test_label_encoder.py::test_empty_input[empty0-ord_label0] distributed.worker - WARNING -  Compute Failed
Function:  _inverse_transform_func
args:      (LabelEncoder(handle_unknown='error'), 0    2
1    1
dtype: int64)
kwargs:    {}
Exception: ValueError('y contains previously unseen label 1')

failures. (https://gpuci.gpuopenanalytics.com/blue/rest/organizations/jenkins/pipelines/rapidsai/pipelines/gpuci/pipelines/cuml/pipelines/prb/pipelines/cuml-gpu-build/runs/15626/log/?start=0)

Rerunning now to see if something transient is going on, otherwise will file as an issue.

[JohnZed]

rerun tests

[JohnZed]

Exception wrapper looks good!

[JohnZed]

rerun tests