v 0.2

[PhilipVinc]

Implement:

• A memory efficient encoding of operators, allowing us to go beyond 20 sites
• Remove Logging utilities, to streamline the dependencies. Those will later be included in a separate package
• Make preliminary changes to Neural Networks initialization so that they can be moved to the GPU
  • In that regard, copy Flux's treelike functions to easily move networks to the gpu.
• Ad a module (undocumented) to compute observables and ground state energy of Closed Hamiltonian systems
• Improve the user-interface, add some methods to dispatch automatically to the correct Steady State Problem given some flags.
• General performance improvements by removing allocations
• Add a lot of unit tests and test on GitHub Actions

Documentation for all this is still missing. Coming soon....