This is Getting Started example to showcase MLEM usage.
Repo has three branches:
main
: produced bydvc repro
, models artifacts are stored with DVCsmall-forest
: we make an experiment withdvc exp
on themain
branch and create this branchno-dvc
: produced bybash run.sh
, this branch has model artifacts pushed straight to Git
You can examine run.sh for pipeline commands or run dvc repro
on dvc.yaml, which has the same outcome as run.sh
, except that with dvc repro
artifacts are stored with DVC.
If you want to reproduce the example from clean state, remove data/
, .mlem/
and scores.json
first.
1. Fork / Clone this repository
git clone [email protected]:iterative/example-mlem.git
cd example-mlem
2. Create virtual environment named venv
python3 -m venv venv
source venv/bin/activate
Install python libraries
pip install --upgrade pip setuptools wheel
pip install -r requirements.txt
3. Run
dvc repro
or:
chmod +x run.sh
./run.sh