More informative error messages for "mixture" efficiency mode #49
Description
Report
If using the mixture model, we should check to ensure that each cell only has one guide instead of throwing this sort of cryptic error during training:
ValueError: Expected parameter probs (Tensor of shape (512, 19212)) of distribution Bernoulli(probs: torch.Size([512, 19212])) to satisfy the constraint Interval(lower_bound=0.0, upper_bound=1.0), but found invalid values:
Version information
No response