❓Question
I want to use aim instead of wandb due to their insane pricing. I want to create a table to track the predictions at each evaluation step where each row is one of the eval examples. It needs to include an image and the other metrics. See below example:
which I implement now in the code like so
table = wandb.Table(
columns=self._val_table_columns,
data=[row for _, row in all_rows],
)
self.logger.experiment.log(
{
"eval/predictions": table,
}
)
is there a way to do this equivalently in aim, if not is this something that could be added?
❓Question
I want to use aim instead of wandb due to their insane pricing. I want to create a table to track the predictions at each evaluation step where each row is one of the eval examples. It needs to include an image and the other metrics. See below example:
which I implement now in the code like so
is there a way to do this equivalently in aim, if not is this something that could be added?