Tom

Actually, it seems that the only verification done on the `value` attribute given to `log_scalar(...)` is in the `MetricsLogger` : https://github.com/IDSIA/sacred/blob/19524bf8b42580c0f30fc8e2a7bcc0001dfbbcb5/sacred/metrics_logger.py#L41-L44 This verification only checks if `value` is a numpy...

You're right, vIsualization can indeed be a problem if the tool you're using expects that each entry of a metric is a scalar (which is a reasonable expectation). But I...

Thanks for your answer! Yes, using a specific `Random` or `RandomState` instance like the one provided by the `_rnd` parameter solves the problem. I agree that this is not a...

Are you using a globally seeded random number generator? If so, your can try to replace it with the [_rnd parameter](https://sacred.readthedocs.io/en/latest/randomness.html#special-arguments) in your captured function as @JarnoRFB suggested. Since the...

Your seeding is correct. The `_rnd` parameter should be used in your code each time you want to use a random operation. For example, replacing all calls to `random.randint` with...

Are you using the `random` module in your code? If so you should use `_rnd` on these calls instead of directly using `random.xxx`. Once its done, PyMongo can do its...