Code for benchmarking AutoGluon.
To get started, run the following commands:
# Do this if you are locally developing AutoGluon to avoid installing it from pip:
git clone https://github.com/autogluon/autogluon
cd autogluon
./full_install.sh
cd ..
Install AutoGluon Bench if running evaluation code.
git clone https://github.com/autogluon/autogluon-bench.git
pip install -e autogluon-bench
Install AutoGluon Benchmark
# Install autogluon-benchmark
git clone https://github.com/Innixma/autogluon-benchmark.git
pip install -e autogluon-benchmark
To run AutoMLBenchmark, see instructions in:
examples/automlbenchmark/README_automlbenchmark.md
Please note that these instructions are quite technical.
To benchmark individual OpenML datasets, you can check out the examples in examples/train_***
The scripts are fairly primitive. They mimic what AutoMLBenchmark does contained to a single dataset.
To generate task metadata files, refer to autogluon_benchmark/data/metadata/README.md