Added - Option for which nonlinearities to use - Option to save models every *n* epochs in training - Option to specify optimization defaults for `e3nn`
Fixed - Using `wandb` no longer breaks the inclusion of special objects like callables in configs
0.3.1
Fixed - `iepoch` is no longer off-by-one when restarting a training run that hit `max_epochs` - Builders, and not just sub-builders, use the class name as a default prefix Added - `early_stopping_xxx` arguments added to enable early stop for platued values or values that out of lower/upper bounds.
0.3.0
Added - Sub-builders can be skipped in `instantiate` by setting them to `None` - More flexible model initialization - Add MD w/ Nequip-ASE-calculator + run-MD script w/ custom Nose-Hoover
Changed - PBC must be explicit if a cell is provided - Training now uses atomic file writes to avoid corruption if interupted - `feature_embedding` renamed to `chemical_embedding` in default models
Fixed - `BesselBasis` now works on GPU when `trainable=False` - Dataset `extra_fixed_fields` are now added even if `get_data()` returns `AtomicData` objects
0.2.1
Fixed - `load_deployed_model` now correctly loads all metadata