How can I change the 7-bit precision limit, it is too small?

Unfortunately, the precision is a critical limit in FHE and can’t be changed: we already have given to our user the maximal available precision (at this time). What you can do however is using quantization, post-training and quantization-aware methods to make that the intermediate values in your model are smaller.

Also, be sure that our research teams work hard to increase this limit. We have already lot of ideas that we need to implement, to have larger precisions. Plus, we’ll have HW accelerations, to make that execution time is less of a problem.

So we hope you can be more patient, and already see what great you can do with the current version.