Integrate ClearML into code that uses autokeras. Initialize a ClearML Task in a code, and ClearML automatically logs scalars, plots, and images reported to TensorBoard, Matplotlib, Plotly, and Seaborn, and all other automatic logging, and explicit reporting added to the code (see Logging).
ClearML allows to:
- Visualize experiment results in the ClearML Web UI.
- Track and upload models.
- Track model performance and create tracking leaderboards.
- Rerun experiments, reproduce experiments on any target machine, and tune experiments.
- Compare experiments.
See the AutoKeras example, which shows ClearML automatically logging:
- The console log
Once these are logged, they can be visualized in the ClearML Web UI.
If you are not already using ClearML, see Getting Started.
Adding ClearML to Code
Add two lines of code:
from clearml import Task
task = Task.init(project_name="myProject", task_name="myExperiment")
When the code runs, it initializes a Task in ClearML Server. A hyperlink to the experiment's log is output to the console.
CLEARML Task: created new task id=c1f1dc6cf2ee4ec88cd1f6184344ca4e
CLEARML results page: https://app.clear.ml/projects/1c7a45633c554b8294fa6dcc3b1f2d4d/experiments/c1f1dc6cf2ee4ec88cd1f6184344ca4e/output/log
Later in the code, define callbacks using TensorBoard, and ClearML logs TensorBoard scalars, histograms, and images.