Quickstart

Get your first fenn experiment running in under five minutes by following these four steps.

Installation

Install the framework via pip:

pip install fenn

Initialize the Project

Fenn includes a CLI to help you jumpstart your project structure. Run this command to download the base template:

fenn pull base

This will generate two essential files:

• fenn.yaml: Your default project configuration and hyperparameters.
• main.py: A boilerplate entrypoint script.

Configure Your Environment

Open fenn.yaml to define your project settings. Fenn separates system logic (logging) from your experiment logic (train).

project: my_project

logger:
  dir: logs

train:
  lr: 0.001

Write and Run the Entrypoint

Fenn uses a decorator-based approach. The @app.entrypoint function receives an args dictionary containing everything defined in your YAML.

Standard Setup

By default, Fenn looks for a file named fenn.yaml.

from fenn import Fenn

app = Fenn()

@app.entrypoint
def main(args):
    # Access config values via the 'args' object
    print(f"Training {args['project']} with LR: {args['train']['lr']}")

if __name__ == "__main__":
    app.run()

Using Custom Config Files

If you want to use a specific configuration (e.g., for different models or environments), use set_config_file() before calling run(). This allows you to swap entire experiment setups instantly.

app = Fenn()
# Point to a specific config file
app.set_config_file("experiments/cnn_baseline.yaml")

@app.entrypoint
def main(args):
    # This will now use settings from cnn_baseline.yaml
    ...

if __name__ == "__main__":
    app.run()

---

The Fenn Execution Lifecycle

When you execute app.run(), Fenn manages the heavy lifting in the background to ensure your experiment is reproducible and organized:

1. Config Selection: It identifies the target YAML file (either fenn.yaml or the file specified via set_config_file).
2. Environment Setup: It automatically creates your logging directory and begins capturing all console output to a timestamped file.
3. Dependency Injection: It starts your main() function, passing the configuration data directly into the args parameter.