log_dict#
- MetricsLogger.log_dict(value_dict: Dict[str, Any], *, key: str | Tuple[str, ...] | None = None, reduce: str | None = 'mean', window: int | float | None = None, ema_coeff: float | None = None, percentiles: List[int] | bool | None = None, clear_on_reduce: bool | None = -1, with_throughput: bool | None = None, throughput_ema_coeff: float | None = -1, reduce_per_index_on_aggregate: bool | None = -1) None[source]#
Logs all leafs of a possibly nested dict of values or Stats objects to this logger.
Traverses through all leafs of
stats_dictand - if a path cannot be found in this logger yet, will add theStatsfound at the leaf under that new key. If a path already exists, will merge the found leaf (Stats) with the ones already logged before. This way,stats_dictdoes NOT have to have the same structure as what has already been logged toself, but can be used to log values under new keys or nested key paths.Passing a dict of stats objects allows you to merge dictionaries of stats objects that have been reduced by other, parallel components.
See MetricsLogger.log_value for more details on the arguments.
- Parameters:
value_dict (Dict[str, Any]) – The possibly nested dict of values or Stats objects to log.
key (str | Tuple[str, ...] | None) – An optional key, or nested key tuple, to prefix the path of each leaf with.
reduce (str | None) – The reduction method. Passed to
log_value()for each leaf.window (int | float | None) – An optional window size. Passed to
log_value()for each leaf.ema_coeff (float | None) – An optional EMA coefficient. Passed to
log_value()for each leaf.percentiles (List[int] | bool | None) – The percentiles to compute. Passed to
log_value()for each leaf.clear_on_reduce (bool | None) – Deprecated. Use
reduce="lifetime_sum"instead.with_throughput (bool | None) – Whether to track a throughput estimate. Passed to
log_value()for each leaf.throughput_ema_coeff (float | None) – Deprecated.
reduce_per_index_on_aggregate (bool | None) – Deprecated.