Get Started: Experiment Tracking

Tools like Jupyter Notebooks are useful for rapid prototyping, but it's hard to keep track of changes and reproduce experiments. You can start using DVC to version your experiments without leaving your Jupyter Notebook. There are no logins, servers, databases, or UI to spin up. Every DVC experiment will be versioned without cluttering your repo, unlike saving each run to a separate directory or creating a Git branch for each.

Running

All you need to start is a DVC repository and the DVCLive Python library installed:

$ pip install dvclive

In your Python code, you can start versioning your experiments in DVCLive's Live API or framework-specific callbacks with save_dvc_exp=True.

There are some examples below (other frameworks available):

Pytorch Lightning

from dvclive import Live
from dvclive.lightning import DVCLiveLogger

...
with Live(save_dvc_exp=True) as live:
    checkpoint = ModelCheckpoint(dirpath="mymodel")
    trainer = Trainer(
        logger=DVCLiveLogger(
            save_dvc_exp=True,
            experiment=live
        ),
        callbacks=checkpoint
    )
    trainer.fit(model)
    live.log_artifact(
        checkpoint.best_model_path,
        type="model",
        name="lightning-model"
    )

Hugging Face

from dvclive import Live
from dvclive.huggingface import DVCLiveCallback

...
with Live(save_dvc_exp=True) as live:
    trainer.add_callback(
        DVCLiveCallback(save_dvc_exp=True, live=live)
    )
    trainer.train()
    trainer.save_model("mymodel")
    live.log_artifact("mymodel", type="model")

Keras

from dvclive import Live
from dvclive.keras import DVCLiveCallback

...
with Live(save_dvc_exp=True) as live:
    model.fit(
        train_dataset,
        validation_data=validation_dataset,
        callbacks=[
            DVCLiveCallback(save_dvc_exp=True, live=live)
        ]
    )
    model.save("mymodel")
    live.log_artifact("mymodel", type="model")

General Python API

from dvclive import Live

with Live(save_dvc_exp=True) as live:
    live.log_param("epochs", NUM_EPOCHS)

    for epoch in range(NUM_EPOCHS):
        train_model(...)
        metrics = evaluate_model(...)
        for metric_name, value in metrics.items():
            live.log_metric(metric_name, value)
        live.next_step()

    live.log_artifact("model.pkl", type="model")

After this, each execution of the code will create a DVC experiment containing the results and the changes needed to reproduce it.

DVCLive will automatically log some metrics, parameters and plots from the ML Framework and any data tracked by DVC but you can also log additional info to be included in the experiment. live.log_artifact("mymodel", type="model") will track your model with DVC and enable managing it with Studio Model Registry.

Learn more about how DVCLive works

Tracking

By following the steps above, you enable different options to monitor the training progress:

DVCLive Report

By default, DVCLive will generate or update a report displaying all the logged data.

If you pass report="notebook" to DVCLive, the report will be displayed and updated inside the output of the cell:

Notebook report

VSCode Extension

The DVC Extension for VS Code will also display all the data logged by DVCLive:

VS Code Report

Studio

If you want to share live updates with others or monitor while away from your machine, follow the instructions in Studio Live Experiments to display updates in the Studio web interface:

Studio Report

Comparing

After you have run multiple experiments, you can compare the results:

DVC CLI

You can use dvc exp show and dvc plots to compare and visualize metrics, parameters and plots across experiments.

$ dvc exp show
─────────────────────────────────────────────────────────────────────────────────────
Experiment                 Created    train.loss   eval.loss   dice_multi   base_lr
─────────────────────────────────────────────────────────────────────────────────────
workspace                  -            0.024942    0.013983        0.922   0.001
master                     05:26 PM      0.78426    0.054157      0.49599   0.1
├── 950c3b5 [bifid-says]   05:33 PM     0.024942    0.013983        0.922   0.001
├── 06090d7 [potty-sash]   05:31 PM     0.026193    0.015237      0.91494   0.01
└── d1ad0a9 [soupy-leak]   05:28 PM     0.075223    0.034786      0.49596   0.1
─────────────────────────────────────────────────────────────────────────────────────

$ dvc plots diff $(dvc exp list --name-only)

plots diff

VSCode Extension

Inside the DVC Extension for VS Code, you can compare and visualize results using the Experiments and Plots views.

Studio

Once you have shared the results to Studio, you can compare experiments against the entire repo history:

Learn more about Comparing Experiments

Get Started: Experiment Tracking

Running

Tracking

Comparing

Content