Philip Hyunsu Cho
Philip Hyunsu Cho
I have been personally maintaining our Jenkins CI server (https://xgboost-ci.net/) since April 2018. While the CI pipeline has been a great asset to the XGBoost project, it suffered from a...
## Location within the Book * Book order number: 11385 * Book build date (Is the date on page 1): September 16, 2020 * Book format (PDF, Epub or Mobi):...
Right now, standalone test executables in `test/` are built with GNU Make ([Makefile](https://github.com/dmlc/dmlc-core/blob/main/test/dmlc_test.mk)). We should also allow these executables to be built with CMake.
https://travis-ci.org/dmlc/dmlc-core/jobs/591522800#L846
It would be nice if every change to dmlc-core is tested against two important downstream projects, XGBoost and MXNet. Let's see if I can get resource for this.
Python 2.x is reaching end-of-life on December 31, 2019. Many projects, including [dmlc/xgboost](https://github.com/dmlc/xgboost) and [apache/incubator-mxnet](https://github.com/apache/incubator-mxnet) have decided to drop support for Python 2. See https://python3statement.org/
Failure may be related to concurrency issue, or not. I need to dive deep and make a diagnosis.
When I compiled dmlc-core with Thread Sanitizer, I got a **massive heap** of concurrency errors: [log.txt](https://github.com/dmlc/dmlc-core/files/3458811/log.txt). This is bad news: multi-threaded code in dmlc-core (threaded iterator, thread groups, concurrent queues...
Some of the outstanding pull requests are more than 90 days old. They should be either merged or closed.
The `normalize=True` option has been [marked as deprecated](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html) in the latest scikit-learn. Use pipelines instead.