Logging and Debug Samples¶
ClearML supports automatic logging and explicit reporting (calls to the ClearML Python Package Logger class methods).
ClearML automatically logs all the following (by adding only two lines of code):
Git repository, branch, commit id, uncommitted changes (git diff), and entry point
Python environment, including specific packages and versions.
Console output from
stderr, and output from external libraries
ArgParser for command line parameters with currently used values.
TensorFlow Definitions (absl-py).
Initial model weights file.
Model snapshots, with optional automatic upload to central storage. Storage options include shared folders, S3, GS, Azure, and Http.
Artifacts logged and stored, including shared folders, S3, GS, Azure, and http(s).
TensorBoard/TensorBoardX, PyTorch, Tensorflow, Keras, AutoKeras, XGBoost, Scikit-Learn
Matplotlib, Plotly, and Seaborn
Resource monitoring (CPU and GPU utilization, memory, video memory, and network usage)
In addition, ClearML supports explicit reporting. Explicitly report the following:
Logging console messages, scalars and plots in several formats, tables, and media including images, audio, and video.
Tracking hyperparameters using parameter dictionaries.
Tracking environment variables.
Storage options for debug samples include: shared folders, S3, GS, Azure, and http(s).
ClearML automatically logs debug samples, allowing you to track and analyze your development process. They are stored in the ClearML Hosted Service (or self-hosted ClearML Server). You can configure ClearML for the storage you use.
The ClearML Python Package contains explicit reporting methods for debug samples include images, audio, video, and HTML.
See the examples for explicitly reporting:
Numpy arrays, uint8, uint8 RGB, and PIL image objects images
Local file, ByteIO stream, and URL image, audio, and video media
Learn about artifacts.