EPIC: Refactor pip package

Current state:

Currently, we offer 3 levels of SSL parts:

• low-level: the separate building blocks such as a projection head, nn memory bank, backbone creation
  https://github.com/lightly-ai/lightly/blob/8dde4d68c0598758bfd05c05f225bf754d9f1ed5/lightly/models/modules/heads.py#L12
• mid-level: the model (backbone + heads + forward functions)
  https://github.com/lightly-ai/lightly/blob/8dde4d68c0598758bfd05c05f225bf754d9f1ed5/lightly/models/moco.py#L13
• high-level: the model, criterion, training loop put together as pytorch lightning module
  https://github.com/lightly-ai/lightly/blob/8dde4d68c0598758bfd05c05f225bf754d9f1ed5/docs/source/getting_started/benchmarks/cifar10_benchmark.py#L144

Target state:

We remove the mid-level representations (delete that code). We provide more examples on the high-level: One example pytorch lightning module for every implemented paper

Steps towards it

Always merge into the development branch: https://github.com/lightly-ai/lightly/tree/master-refactor-SSL-models

First step:

• #494
• https://github.com/lightly-ai/lightly/issues/492
• #498
• https://github.com/lightly-ai/lightly/issues/505
• #508
• #512

Second step:

• #515
• Remove mid- to high-level code from CLI
• https://github.com/lightly-ai/lightly/issues/380

Third step:

• delete mid-level models in lightly/lightly/models

Start by reviewing the mid-level model implementations under lightly/lightly/models and compare them with the low-level modules and high-level benchmark example linked in the issue. Map the remaining mid-level models and CLI references, then verify that each implemented paper has a corresponding high-level PyTorch Lightning example and that the mid-level model code is removed without breaking tests.