Get Started: Metrics, Parameters, and Plots

🎬 Click to watch a video intro.

DVC makes it easy to track metrics, update parameters, and visualize performance with plots. These concepts are introduced below.

All of the above can be combined into experiments to run and compare many iterations of your ML project.

Collecting metrics

First, let's see the mechanism to capture values for these ML attributes. Add and run a final evaluation stage to our earlier pipeline:

$ dvc stage add -n evaluate \
  -d src/evaluate.py -d model.pkl -d data/features \
  -M eval/live/metrics.json -O eval/live/plots \
  -O eval/prc -o eval/importance.png \
  python src/evaluate.py model.pkl data/features

$ dvc repro

💡 Expand to see what happens under the hood.

The -O option here specifies an output that will not be cached by DVC, and -M specifies a metrics file (that will also not be cached). dvc stage add generates a new stage in the dvc.yaml file:

evaluate:
  cmd: python src/evaluate.py model.pkl data/features
  deps:
    - data/features
    - model.pkl
    - src/evaluate.py
  outs:
    - eval/importance.png
    - eval/live/plots:
        cache: false
    - eval/prc:
        cache: false
  metrics:
    - eval/live/metrics.json:
        cache: false

The biggest difference from previous stages in our pipeline is the new metrics section. Metrics files contain scalar values (e.g. AUC) to compare across iterations.

With cache: false, DVC skips caching the output, as we want these JSON metrics files to be versioned by Git.

evaluate.py writes the model's ROC-AUC and average precision to eval/live/metrics.json (previously marked as a metrics file with -M):

{
  "avg_prec": {
    "train": 0.9772271756725741,
    "test": 0.9449556493816984
  },
  "roc_auc": {
    "train": 0.9873675866013153,
    "test": 0.9619097316125981
  }
}

You can view tracked metrics with dvc metrics show :

$ dvc metrics show
Path                    avg_prec.test    avg_prec.train    roc_auc.test    roc_auc.train
eval/live/metrics.json  0.94496          0.97723           0.96191         0.98737

Visualizing plots

The stage also writes different files with data that can be graphed:

DVCLive-generated roc_curve and confusion_matrix values in the eval/live/plots directory.

Precision-recall curves as JSON arrays in eval/prc/train.json:

{
  "prc": [
    { "precision": 0.0215, "recall": 1.0, "threshold": 0.0 },
    { "precision": 1.0, "recall": 0.0093, "threshold": 0.6 },
    ...

A custom eval/importance.png image showing a bar chart of features' importance.

You can visualize all of these with DVC! Start by configuring the plots in dvc.yaml:

plots:
  - ROC:
      x: fpr
      y:
        eval/live/plots/sklearn/roc/train.json: tpr
        eval/live/plots/sklearn/roc/test.json: tpr
  - Confusion-Matrix:
      template: confusion
      x: actual
      y:
        eval/live/plots/sklearn/cm/train.json: predicted
        eval/live/plots/sklearn/cm/test.json: predicted
  - Precision-Recall:
      x: recall
      y:
        eval/prc/train.json: precision
        eval/prc/test.json: precision
  - eval/importance.png

To generate them, you can run dvc plots show (shown below), which generates an HTML file you can open in a browser. Or you can load your project in VS Code and use the DVC Extension's Plots Dashboard.

$ dvc plots show
file:///Users/dvc/example-get-started/dvc_plots/index.html

Let's save this iteration so we can compare it later:

$ git add .gitignore dvc.yaml dvc.lock eval
$ git commit -a -m "Create evaluation stage"

Later we will see how to compare and visualize different pipeline iterations. For now, let's see how to capture another important piece of information which will be useful for comparison: parameters.

Defining stage parameters

It's pretty common for data science pipelines to include configuration files that define adjustable parameters to train a model, do pre-processing, etc. DVC provides a mechanism for stages to depend on the values of specific sections of such a config file (YAML, JSON, TOML, and Python formats are supported).

Luckily, we should already have a stage with parameters in dvc.yaml:

featurize:
  cmd: python src/featurization.py data/prepared data/features
  deps:
    - data/prepared
    - src/featurization.py
  params:
    - featurize.max_features
    - featurize.ngrams
  outs:
    - data/features

⚙️ Expand to recall how it was generated.

The featurize stage was created with this dvc run command. Notice the argument sent to the -p option (short for --params):

$ dvc run -n featurize \
          -p featurize.max_features,featurize.ngrams \
          -d src/featurization.py -d data/prepared \
          -o data/features \
          python src/featurization.py data/prepared data/features

The params section defines the parameter dependencies of the featurize stage. By default, DVC reads those values (featurize.max_features and featurize.ngrams) from a params.yaml file. But as with metrics and plots, parameter file names and structure can also be user- and case-defined.

Here's the contents of our params.yaml file:

prepare:
  split: 0.20
  seed: 20170428

featurize:
  max_features: 100
  ngrams: 1

train:
  seed: 20170428
  n_est: 50
  min_split: 2

Updating params and iterating

We are definitely not happy with the AUC value we got so far! Let's edit the params.yaml file to use bigrams and increase the number of features:

 featurize:
-  max_features: 100
-  ngrams: 1
+  max_features: 200
+  ngrams: 2

The beauty of dvc.yaml is that all you need to do now is run:

$ dvc repro

It'll analyze the changes, use existing results from the run-cache, and execute only the commands needed to produce new results (model, metrics, plots).

The same logic applies to other possible adjustments — edit source code, update datasets — you do the changes, use dvc repro, and DVC runs what needs to be run.

Comparing iterations

Finally, let's see how the updates improved performance. DVC has a few commands to see changes in and visualize metrics, parameters, and plots. These commands can work for one or across multiple pipeline iteration(s). Let's compare the current "bigrams" run with the last committed "baseline" iteration:

$ dvc params diff
Path         Param                   HEAD  workspace
params.yaml  featurize.max_features  100   200
params.yaml  featurize.ngrams        1     2

dvc params diff can show how params in the workspace differ vs. the last commit.

dvc metrics diff does the same for metrics:

$ dvc metrics diff
Path                    Metric          HEAD     workspace    Change
eval/live/metrics.json  avg_prec.test   0.9014   0.925        0.0236
eval/live/metrics.json  avg_prec.train  0.95704  0.97437      0.01733
eval/live/metrics.json  roc_auc.test    0.93196  0.94602      0.01406
eval/live/metrics.json  roc_auc.train   0.97743  0.98667      0.00924

And finally, we can compare all plots with a single command (we show only some of them for simplicity):

$ dvc plots diff
file:///Users/dvc/example-get-started/plots.html

See dvc plots diff for more info on its options.

All these commands also accept Git revisions (commits, tags, branch names) to compare.

On the next page, you can learn advanced ways to track, organize, and compare more experiment iterations.