tracker-ml documentation

Contents:

• Installation
  • Requirements
• CLI Tool
  • Examples
• SDK

tracker-ml is the SDK and command line interface used with www.tracker.ml. These tools enable local file tracking for every run of a machine learning project. Along with files, parameters can also be tracked so that each run can be compared.

Example

Initialize tracker using the CLI and add the files you want to track. Then import tracker_ml.tml anywhere and everywhere in your project. Use is easy:

import tracker_ml.tml as tml

tml.login("username", "password")
tml.model("Logistic Regression")
tml.record("description", "Tracker setup")

# <machine learning code>

# record int, float, or str
tml.accuracy(0.42)
tml.record("alpha", 0.9)

# record multiple values under the same key
tml.mrecord("epoch", 1)
tml.mrecord("epoch", 2)
tml.mrecord("epoch", 3)