Quick Start¶

This guide will help you get started with ExpOps using the built-in template.

Step 1: Create a Workspace¶

First, create a workspace directory for your ExpOps projects:

mkdir -p ~/expops-workspace && cd ~/expops-workspace

Step 2: Create a Project from Template¶

Use the sklearn-basic template to create your first project:

expops create sklearn-basic --template sklearn-basic

This creates a new project at sklearn-basic/ with: - Configuration files - Model code - Chart generation scripts - Requirements files - Example data structure

Step 3: Run the Project¶

Run the project locally:

expops run sklearn-basic --local

This will:

Set up the virtual environment
Install dependencies
Execute the pipeline
Generate artifacts and charts

Step 4: View Results¶

The project configuration is located at:

sklearn-basic/configs/project_config.yaml

By default, the template uses a local (SQLite) cache backend, which is persistent. To enable cross-process live metrics (web UI) or remote backends, update experiment.parameters.cache in the project config.

Important: Caching and Web UI Requirements¶

The default cache.backend.type: local supports persistent caching. For web UI metrics and charts across runs, or for remote/shared setups, set cache.backend.type to gcp (Firestore) and optionally configure cache.object_store for GCS. Example in project_config.yaml:

experiment:
  parameters:
    cache:
      backend:
        type: local   # default; use gcp for Firestore (web UI / shared runs)
        # For GCP (Firestore):
        # type: gcp
        # gcp_project: <your-gcp-project-id>
        # credentials_json: firestore.json
      # Optional object store for cache artifacts (e.g. GCS):
      # object_store:
      #   type: gcs
      #   bucket: <your-gcs-bucket-name>
      #   # Manifests will be stored under gs://<bucket>/<project_id>/cache/steps/ automatically.

Setup steps for GCP (Firestore): 1. Create a Firestore database in Google Cloud 2. Add credentials to sklearn-basic/firestore.json 3. Set cache.backend.type to gcp and set gcp_project and credentials_json under cache.backend

See the Backends documentation for more details and setup instructions.

Running on a Cluster¶

To run on a distributed cluster (e.g., SLURM):

Add a compute_config.yaml under configs/
Remove the --local flag when running:

expops run sklearn-basic