๐Ÿ“Š Evaluation#

To track training progress, validation datasets can be provided to the Composer Trainer through the eval_dataloader parameter. The trainer will compute evaluation metrics on the evaluation dataset at a frequency specified by the the Trainer parameter eval_interval.

from composer import Trainer

trainer = Trainer(
    eval_interval="1ep",  # Default is every epoch

The metrics should be provided by ComposerModel.metrics(). For more information, see the โ€œMetricsโ€ section in ๐Ÿ›ป ComposerModel.

To provide a deeper intuition, hereโ€™s pseudocode for the evaluation logic that occurs every eval_interval:

metrics = model.metrics(train=False)

for batch in eval_dataloader:
    outputs, targets = model.validate(batch)
    metrics.update(outputs, targets)  # implements the torchmetrics interface

  • The trainer iterates over eval_dataloader and passes each batch to the modelโ€™s ComposerModel.validate() method.

  • Outputs of model.validate are used to update metrics (a torchmetrics.Metric or torchmetrics.MetricCollection returned by .ComposerModel.metrics).

  • Finally, metrics over the whole validation dataset are computed.

Note that the tuple returned by ComposerModel.validate() provide the positional arguments to metrics.update. Please keep this in mind when using custom models and/or metrics.

Multiple Datasets#

If there are multiple validation datasets that may have different metrics, use Evaluator to specify each pair of dataloader and metrics. This class is just a container for a few attributes:

  • label: a user-specified name for the evaluator.

  • dataloader: PyTorch DataLoader or our DataSpec.

    See DataLoaders for more details.

  • metric_names: list of names of metrics to track.

For example, the GLUE tasks for language models can be specified as in the following example:

from composer.core import Evaluator
from composer.models.nlp_metrics import BinaryF1Score

glue_mrpc_task = Evaluator(
    metric_names=['BinaryF1Score', 'Accuracy']

glue_mnli_task = Evaluator(

trainer = Trainer(
    eval_dataloader=[glue_mrpc_task, glue_mnli_task],

Note that metric_names must be a subset of the metrics provided by the model in ComposerModel.metrics().