Metric logger for ML projects.
Project description
DVCLive
DVCLive is an open-source library for monitoring machine learning model performance. It’s an ML logger similar to MLFlow, Weights & Biases, Neptune, Tensorboard, etc., but built on top of DVC, and with Git and MLOps principles in mind:
Codification of data. Tracked metrics are stored in readable text files that can be versioned by Git or other version control tools.
Distributed. No services or servers are required. Metrics are stored in a Git repository as text files, or pointers to files in DVC storage.
GitOps API. Plots are generated through DVC using Git commit SHAs or branch names, e.g.:
dvc plots diff --target logs master
.
Automation. DVCLive metrics are easy to use by any automation, DevOps, or MLOps tool such as CI/CD (including CML), custom scripts, or ML platforms.
Python API
DVCLive is a Python library. The interface consists of three main methods:
dvclive.init(path)
- initializes a DVCLive logger. The metrics will be saved underpath
.dvclive.log(metric, value, step)
- logs the metric value. Thevalue
andstep
(optional) will be appended topath/{metric}.tsv
file.dvclive.next_step()
- signals that the current step has ended (implied when the samemetric
is logged again).
Call to collaboration
Today only Python is supported (while DVC is language agnostic), with a minimum number of connectors to ML libs (Keras, XGBoost). The DVCLive team is happy to extend the functionality as needed. Please create an issue to start a discussion!
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for dvclive-0.0.3-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 89c3f2fc89be7fe46cda7f3d328d06029481b46891e1580fd74ca36cf1828f62 |
|
MD5 | fa8132595f847ff1c23189cb231c0a56 |
|
BLAKE2b-256 | 0bdbf6bf13b8da834a16deef166b7172ff48a4c62ce41ebd53b9f74ac6b479cd |