eschibli
eschibli
Thanks for the suggestion haifeng, that should help in the meantime.
@haifeng-jin @elloza I did try, and as elloza says, it isn't working. Varying the tuner does not appear to prevent the same model from being repeatedly tried and rejected for...
@haifeng-jin Without looking at the current code, it seems like simply assigning very high or `Inf` loss to invalid models would solve the problem rather elegently.
I can't vouch for whether it's performing correctly or not, but setting the batch size to the lookback allows it to run for me.
This is [Issue 1554](https://github.com/keras-team/autokeras/issues/1554). It seems possible to work around it by setting the batch size so that `lookback` is divisible by `batch_size`.
I would find this helpful as well
@rzese this is [issue 1479](https://github.com/keras-team/autokeras/issues/1479). It hasn't been solved yet and I'm not aware of a workaround unfortunately.
> The issue is that minkowski with p< 1 is not a metric (no triangular inequality). Then the question is do we want to support similarity measures that are not...
I will try to double-check this later this week.
The issue is that the column is truncated - it should read `-1.826E+05`, but the five is cut off. I'm not sure what causes this because I'm not an expert...