# "classic" TPS with openpathsampling on 2D toy potentials

```{toctree}
:maxdepth: 1
:caption: 2D toy potential notebooks

1 toy pytorch simple: setup <1_Toy_pytorch_simple_setup>
1 toy pytorch simple: setup XY diagpotential (Hummer-Szabo potential) <1_Toy_pytorch_simple_setup_XYDiagpot>
2 toy pytorch simple: continuing simulations <2_Toy_pytorch_continuing_simulations>
3 toy symbolic regression <3_Toy_SymbolicRegression>
4 toy: loading and saving rcmodels <4_Toy_loading_and_saving_rcmodels>

1.5 toy pytorch multidomain model: setup <1.5_Toy_pytorch_MultiDomainModel_setup>

1 toy tensorflow simple: setup <1_Toy_tensorflow_simple_setup>
```

This folder contains a number of notebooks using aimmd to learn the committor of 2D toy potentials (with added orthogonal and irrelevant dimensions to make the learning task more challenging and realistic).
There are various variants of these notebooks using different neural network architectures and machine learning backends, they do not differ in any other regard, so it is probably enough to do one "sequence" of notebooks.