This directory contains pre-generated data. For example, for the rs256 example token we'll find

[4.0K]  rs256                           // Each example token has it's own directory
├── [4.0K]  cache                       // Cached output of the one-time operations that are slow
│   ├── [ 39K]  client_state.bin        // This contains the user's proof, generated once per token
│   └── [580M]  groth16_params.bin      // These are the Groth16 circuit-specific system parameters 
                                        // (common to all users with a particular token type)
├── [ 249]  io_locations.sym            // Encodes the positions of the outputs of the Groth16 circuit
├── [595M]  main_c.r1cs                 // The R1CS instance associated with the Groth16 circuit
├── [6.3M]  main.wasm                   // WASM for witness generation (used by the Groth16 prover)
├── [  36]  prover_aux.json             // Auxiliary info the prover needs to create presentation proofs
├── [ 18K]  prover_inputs.json          // Inputs known to the prover for the Groth16 circuit
└── [ 818]  public_IOs.json             // Public inputs/outputs for the Groth16 circuit

After running one of the examples for the first time the cache will be populated. On subsequent runs it'll be used.

The other data all comes from the code in crescent/setup.