# DVCLive
[](https://pypi.org/project/dvclive/)
[](https://pypi.org/project/dvclive/)
[](https://pypi.org/project/dvclive)
[](https://opensource.org/licenses/Apache-2.0)
[](https://github.com/iterative/dvclive/actions?workflow=Tests)
[](https://app.codecov.io/gh/iterative/dvclive)
[](https://github.com/pre-commit/pre-commit)
[](https://github.com/psf/black)
DVCLive is a Python library for logging machine learning metrics and other
metadata in simple file formats, which is fully compatible with DVC.
# [Documentation](https://dvc.org/doc/dvclive)
- [Get Started](https://dvc.org/doc/start/experiments)
- [How it Works](https://dvc.org/doc/dvclive/how-it-works)
- [API Reference](https://dvc.org/doc/dvclive/live)
- [Integrations](https://dvc.org/doc/dvclive/ml-frameworks)
______________________________________________________________________
# Quickstart
| Python API Overview | PyTorch Lightning | Scikit-learn | Ultralytics YOLO v8 |
|--------|--------|--------|--------|
| <a href="https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-Quickstart.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" /></a> | <a href="https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-PyTorch-Lightning.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" /></a> | <a href="https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-scikit-learn.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" /></a> | <a href="https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-YOLO.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" /></a> |
## Install *dvclive*
```console
$ pip install dvclive
```
## Initialize DVC Repository
```console
$ git init
$ dvc init
$ git commit -m "DVC init"
```
## Example code
Copy the snippet below into `train.py` for a basic API usage example:
```python
import time
import random
from dvclive import Live
params = {"learning_rate": 0.002, "optimizer": "Adam", "epochs": 20}
with Live() as live:
# log a parameters
for param in params:
live.log_param(param, params[param])
# simulate training
offset = random.uniform(0.2, 0.1)
for epoch in range(1, params["epochs"]):
fuzz = random.uniform(0.01, 0.1)
accuracy = 1 - (2 ** - epoch) - fuzz - offset
loss = (2 ** - epoch) + fuzz + offset
# log metrics to studio
live.log_metric("accuracy", accuracy)
live.log_metric("loss", loss)
live.next_step()
time.sleep(0.2)
```
See [Integrations](https://dvc.org/doc/dvclive/ml-frameworks) for examples using
DVCLive alongside different ML Frameworks.
## Running
Run this a couple of times to simulate multiple experiments:
```console
$ python train.py
$ python train.py
$ python train.py
...
```
## Comparing
DVCLive outputs can be rendered in different ways:
### DVC CLI
You can use [dvc exp show](https://dvc.org/doc/command-reference/exp/show) and
[dvc plots](https://dvc.org/doc/command-reference/plots) to compare and
visualize metrics, parameters and plots across experiments:
```console
$ dvc exp show
```
```
─────────────────────────────────────────────────────────────────────────────────────────────────────────────
Experiment Created train.accuracy train.loss val.accuracy val.loss step epochs
─────────────────────────────────────────────────────────────────────────────────────────────────────────────
workspace - 6.0109 0.23311 6.062 0.24321 6 7
master 08:50 PM - - - - - -
├── 4475845 [aulic-chiv] 08:56 PM 6.0109 0.23311 6.062 0.24321 6 7
├── 7d4cef7 [yarer-tods] 08:56 PM 4.8551 0.82012 4.5555 0.033533 4 5
└── d503f8e [curst-chad] 08:56 PM 4.9768 0.070585 4.0773 0.46639 4 5
─────────────────────────────────────────────────────────────────────────────────────────────────────────────
```
```console
$ dvc plots diff $(dvc exp list --names-only) --open
```
### DVC Extension for VS Code
Inside the
[DVC Extension for VS Code](https://marketplace.visualstudio.com/items?itemName=Iterative.dvc),
you can compare and visualize results using the
[Experiments](https://github.com/iterative/vscode-dvc/blob/main/extension/resources/walkthrough/experiments-table.md)
and
[Plots](https://github.com/iterative/vscode-dvc/blob/main/extension/resources/walkthrough/plots.md)
views:
While experiments are running, live updates will be displayed in both views.
### DVC Studio
If you push the results to [DVC Studio](https://dvc.org/doc/studio), you can
compare experiments against the entire repo history:
You can enable
[Studio Live Experiments](https://dvc.org/doc/studio/user-guide/projects-and-experiments/live-metrics-and-plots)
to see live updates while experiments are running.
______________________________________________________________________
# Comparison to related technologies
**DVCLive** is an *ML Logger*, similar to:
- [MLFlow](https://mlflow.org/)
- [Weights & Biases](https://wandb.ai/site)
- [Neptune](https://neptune.ai/)
The main differences with those *ML Loggers* are:
- **DVCLive** does not **require** any additional services or servers to run.
- **DVCLive** metrics, parameters, and plots are
[stored as plain text files](https://dvc.org/doc/dvclive/how-it-works#directory-structure)
that can be versioned by tools like Git or tracked as pointers to files in DVC
storage.
- **DVCLive** can save experiments or runs as
[hidden Git commits](https://dvc.org/doc/dvclive/how-it-works#track-the-results).
You can then use different [options](#comparing) to visualize the metrics,
parameters, and plots across experiments.
______________________________________________________________________
# Contributing
Contributions are very welcome. To learn more, see the
[Contributor Guide](CONTRIBUTING.rst).
# License
Distributed under the terms of the
[Apache 2.0 license](https://opensource.org/licenses/Apache-2.0), *dvclive* is
free and open source software.
Raw data
{
"_id": null,
"home_page": null,
"name": "dvclive",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": "Iterative <support@dvc.org>",
"keywords": "ai, metrics, collaboration, data-science, data-version-control, developer-tools, git, machine-learning, reproducibility",
"author": null,
"author_email": "Iterative <support@dvc.org>",
"download_url": "https://files.pythonhosted.org/packages/8d/7c/df9ef47ed816c9c31c36fb0f8d2d4cceaf3b5ae9da2c3bf0babb74d52d58/dvclive-3.48.1.tar.gz",
"platform": "any",
"description": "# DVCLive\n\n[](https://pypi.org/project/dvclive/)\n[](https://pypi.org/project/dvclive/)\n[](https://pypi.org/project/dvclive)\n[](https://opensource.org/licenses/Apache-2.0)\n\n[](https://github.com/iterative/dvclive/actions?workflow=Tests)\n[](https://app.codecov.io/gh/iterative/dvclive)\n[](https://github.com/pre-commit/pre-commit)\n[](https://github.com/psf/black)\n\nDVCLive is a Python library for logging machine learning metrics and other\nmetadata in simple file formats, which is fully compatible with DVC.\n\n# [Documentation](https://dvc.org/doc/dvclive)\n\n- [Get Started](https://dvc.org/doc/start/experiments)\n- [How it Works](https://dvc.org/doc/dvclive/how-it-works)\n- [API Reference](https://dvc.org/doc/dvclive/live)\n- [Integrations](https://dvc.org/doc/dvclive/ml-frameworks)\n\n______________________________________________________________________\n\n# Quickstart\n\n| Python API Overview | PyTorch Lightning | Scikit-learn | Ultralytics YOLO v8 |\n|--------|--------|--------|--------|\n| <a href=\"https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-Quickstart.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" /></a> | <a href=\"https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-PyTorch-Lightning.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" /></a> | <a href=\"https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-scikit-learn.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" /></a> | <a href=\"https://colab.research.google.com/github/iterative/dvclive/blob/main/examples/DVCLive-YOLO.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" /></a> |\n\n## Install *dvclive*\n\n```console\n$ pip install dvclive\n```\n\n## Initialize DVC Repository\n\n```console\n$ git init\n$ dvc init\n$ git commit -m \"DVC init\"\n```\n\n## Example code\n\nCopy the snippet below into `train.py` for a basic API usage example:\n\n```python\nimport time\nimport random\n\nfrom dvclive import Live\n\nparams = {\"learning_rate\": 0.002, \"optimizer\": \"Adam\", \"epochs\": 20}\n\nwith Live() as live:\n\n # log a parameters\n for param in params:\n live.log_param(param, params[param])\n\n # simulate training\n offset = random.uniform(0.2, 0.1)\n for epoch in range(1, params[\"epochs\"]):\n fuzz = random.uniform(0.01, 0.1)\n accuracy = 1 - (2 ** - epoch) - fuzz - offset\n loss = (2 ** - epoch) + fuzz + offset\n\n # log metrics to studio\n live.log_metric(\"accuracy\", accuracy)\n live.log_metric(\"loss\", loss)\n live.next_step()\n time.sleep(0.2)\n```\n\nSee [Integrations](https://dvc.org/doc/dvclive/ml-frameworks) for examples using\nDVCLive alongside different ML Frameworks.\n\n## Running\n\nRun this a couple of times to simulate multiple experiments:\n\n```console\n$ python train.py\n$ python train.py\n$ python train.py\n...\n```\n\n## Comparing\n\nDVCLive outputs can be rendered in different ways:\n\n### DVC CLI\n\nYou can use [dvc exp show](https://dvc.org/doc/command-reference/exp/show) and\n[dvc plots](https://dvc.org/doc/command-reference/plots) to compare and\nvisualize metrics, parameters and plots across experiments:\n\n```console\n$ dvc exp show\n```\n\n```\n\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\nExperiment Created train.accuracy train.loss val.accuracy val.loss step epochs\n\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\nworkspace - 6.0109 0.23311 6.062 0.24321 6 7\nmaster 08:50 PM - - - - - -\n\u251c\u2500\u2500 4475845 [aulic-chiv] 08:56 PM 6.0109 0.23311 6.062 0.24321 6 7\n\u251c\u2500\u2500 7d4cef7 [yarer-tods] 08:56 PM 4.8551 0.82012 4.5555 0.033533 4 5\n\u2514\u2500\u2500 d503f8e [curst-chad] 08:56 PM 4.9768 0.070585 4.0773 0.46639 4 5\n\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\n```\n\n```console\n$ dvc plots diff $(dvc exp list --names-only) --open\n```\n\n\n\n### DVC Extension for VS Code\n\nInside the\n[DVC Extension for VS Code](https://marketplace.visualstudio.com/items?itemName=Iterative.dvc),\nyou can compare and visualize results using the\n[Experiments](https://github.com/iterative/vscode-dvc/blob/main/extension/resources/walkthrough/experiments-table.md)\nand\n[Plots](https://github.com/iterative/vscode-dvc/blob/main/extension/resources/walkthrough/plots.md)\nviews:\n\n\n\n\n\nWhile experiments are running, live updates will be displayed in both views.\n\n### DVC Studio\n\nIf you push the results to [DVC Studio](https://dvc.org/doc/studio), you can\ncompare experiments against the entire repo history:\n\n\n\nYou can enable\n[Studio Live Experiments](https://dvc.org/doc/studio/user-guide/projects-and-experiments/live-metrics-and-plots)\nto see live updates while experiments are running.\n\n______________________________________________________________________\n\n# Comparison to related technologies\n\n**DVCLive** is an *ML Logger*, similar to:\n\n- [MLFlow](https://mlflow.org/)\n- [Weights & Biases](https://wandb.ai/site)\n- [Neptune](https://neptune.ai/)\n\nThe main differences with those *ML Loggers* are:\n\n- **DVCLive** does not **require** any additional services or servers to run.\n- **DVCLive** metrics, parameters, and plots are\n [stored as plain text files](https://dvc.org/doc/dvclive/how-it-works#directory-structure)\n that can be versioned by tools like Git or tracked as pointers to files in DVC\n storage.\n- **DVCLive** can save experiments or runs as\n [hidden Git commits](https://dvc.org/doc/dvclive/how-it-works#track-the-results).\n\nYou can then use different [options](#comparing) to visualize the metrics,\nparameters, and plots across experiments.\n\n______________________________________________________________________\n\n# Contributing\n\nContributions are very welcome. To learn more, see the\n[Contributor Guide](CONTRIBUTING.rst).\n\n# License\n\nDistributed under the terms of the\n[Apache 2.0 license](https://opensource.org/licenses/Apache-2.0), *dvclive* is\nfree and open source software.\n",
"bugtrack_url": null,
"license": "Apache License 2.0",
"summary": "Experiments logger for ML projects.",
"version": "3.48.1",
"project_urls": {
"Changelog": "https://github.com/iterative/dvclive/releases",
"Documentation": "https://dvc.org/doc/dvclive",
"Homepage": "https://github.com/iterative/dvclive",
"Issues": "https://github.com/iterative/dvclive/issues",
"Repository": "https://github.com/iterative/dvclive"
},
"split_keywords": [
"ai",
" metrics",
" collaboration",
" data-science",
" data-version-control",
" developer-tools",
" git",
" machine-learning",
" reproducibility"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "3819774a835847bf497ac2d0bfdd40c5332760a5ad63dbf44c1851a22b1a2cad",
"md5": "1477d14cbd8cc89a4c6507577134556d",
"sha256": "d1fdc14fe9c0b5a50609edbd813c2b681daabd48008605f74b1eee2af325549e"
},
"downloads": -1,
"filename": "dvclive-3.48.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "1477d14cbd8cc89a4c6507577134556d",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 43559,
"upload_time": "2024-12-19T06:33:49",
"upload_time_iso_8601": "2024-12-19T06:33:49.912665Z",
"url": "https://files.pythonhosted.org/packages/38/19/774a835847bf497ac2d0bfdd40c5332760a5ad63dbf44c1851a22b1a2cad/dvclive-3.48.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "8d7cdf9ef47ed816c9c31c36fb0f8d2d4cceaf3b5ae9da2c3bf0babb74d52d58",
"md5": "1fcfc73e93714b5f33c01b8792e30aec",
"sha256": "1aaa0943026fe115f59bb7aaf0219197c67e6673805126003d2c0210e7e59299"
},
"downloads": -1,
"filename": "dvclive-3.48.1.tar.gz",
"has_sig": false,
"md5_digest": "1fcfc73e93714b5f33c01b8792e30aec",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 2041931,
"upload_time": "2024-12-19T06:33:52",
"upload_time_iso_8601": "2024-12-19T06:33:52.795797Z",
"url": "https://files.pythonhosted.org/packages/8d/7c/df9ef47ed816c9c31c36fb0f8d2d4cceaf3b5ae9da2c3bf0babb74d52d58/dvclive-3.48.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-19 06:33:52",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "iterative",
"github_project": "dvclive",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "dvclive"
}
JSON
