`nemo_automodel.components.loggers.comet_utils`#

Module Contents#

Classes#

CometLogger

Comet ML logger for experiment tracking.

Functions#

build_comet

Build Comet logger from configuration.

Data#

logger

API#

nemo_automodel.components.loggers.comet_utils.logger#: ‘getLogger(…)’

class nemo_automodel.components.loggers.comet_utils.CometLogger(

project_name: str,

workspace: Optional[str] = None,

api_key: Optional[str] = None,

experiment_name: Optional[str] = None,

tags: Optional[list] = None,

auto_metric_logging: bool = False,

**kwargs,

)#

Comet ML logger for experiment tracking.

Initialization

Initialize Comet ML logger.

Parameters:

project_name – Name of the Comet project
workspace – Comet workspace (optional, uses default from config/env)
api_key – Comet API key (optional, uses COMET_API_KEY env var)
experiment_name – Name for this experiment run (optional)
tags – List of tags to add to the experiment
auto_metric_logging – Whether to enable Comet’s auto metric logging
**kwargs – Additional arguments passed to comet_ml.Experiment()

log_params(params: Dict[str, Any]) → None#

Log parameters to Comet.

Parameters:: params – Dictionary of parameters to log

log_metrics( metrics: Dict[str, float], step: Optional[int] = None, ) → None#

Log metrics to Comet.