Documentation
BlogsCreate Account
Start typing to search…

Quick Start

To get started, we need to have the mlfoundry library installed. You can install it following the instructions in the CLI Setup docs.

Create a ML Repo

📘

Prerequiste - Blob Storage Integration

Before you can create an ML Repo, you'd need to connect one or more Blob Storages (S3, GCS, Azure Blob, MinIO, etc) to store artifacts and models associated with a ML Repo. If this one time setup is already done, you can skip to next section

You can refer to one of the following pages to connect your blob storage to TrueFoundry

You can then create an ML Repo from the ML Repo's tab in the Platform.


Create a run and start tracking

Initialize a new run. This run will now appear on the TrueFoundry dashboard under ML Repos Tab.

from truefoundry.ml import get_client

client = get_client()
run = client.create_run(ml_repo="iris-demo", run_name="svm-model")

Track parameters

Save hyperparameters for your experiment.

run.log_params({"learning_rate": 0.001})

Track metrics

Save metrics for your model.

run.log_metrics({"accuracy": 0.7, "loss": 0.6})

End a run and stop tracking

After completion of your experiment, you can end a run. This function marks the run as “finished” and stops tracking system metrics.

run.end()

Congratulations! You have successfully created a new ML repository. It is now ready for use, and you can start populating it with your data, code, models, and other related resources.

Login without manual interaction ( Non-Interactive Mode )

TrueFoundry offers a convenient option to automate login through the command-line interface (CLI) for integration within your scripts. This non-interactive approach utilizes environment variables, eliminating the need for manual input (approving through clicking the Approve button in the browser that opens)

In non-interactive mode, you can set environment variables to automate the login process. To do this, set the following environment variables:

Updated about 1 month ago

What’s Next