Ctrl+K

Benchmarking

Benchmarking#

Benchmarking helper functions.

anomalib.pipelines.benchmarking.distribute(config_path)#

Run all cpu experiments on a single process. Distribute gpu experiments over all available gpus.

Parameters:

config_path (Path) – (Path): Config path.

Return type:

None

anomalib.pipelines.benchmarking.upload_to_comet(folder=None)#

Upload the data in csv files to comet.

Creates a project named benchmarking_[two random characters]. This is so that the project names are unique. One issue is that it does not check for collision

Parameters:

folder (optional, str) – Sub-directory from which runs are picked up. Defaults to None. If none picks from runs.

Return type:

None

anomalib.pipelines.benchmarking.upload_to_wandb(team='anomalib', folder=None)#

Upload the data in csv files to wandb.

Creates a project named benchmarking_[two random characters]. This is so that the project names are unique. One issue is that it does not check for collision

Parameters:

  • team (str, optional) – Name of the team on wandb. This can also be the id of your personal account.

  • "anomalib". (Defaults to) –

  • folder (optional, str) – Sub-directory from which runs are picked up. Defaults to None. If none picks from runs.

Return type:

None

anomalib.pipelines.benchmarking.write_metrics(model_metrics, writers, folder=None)#

Write metrics to destination provided in the sweep config.

Parameters:

  • model_metrics (dict) – Dictionary to be written

  • writers (list[str]) – List of destinations.

  • folder (optional, str) – Sub-directory to which runs are written to. Defaults to None. If none writes to root.

Return type:

None

Contents