This folder contains:
```python
all_circuit_parameters_train.pickle
all_simulation_outs_train.pickle
all_seeds_train.npy
```

The first two are pandas dataframes which contain the `valid` theta and x from **all** 
rounds.

In addition, it contains folder:
```python
01_simulate_11deg
01_simulate_11deg_R2
01_simulate_11deg_R3
```
These contain the raw simulations from all three rounds of rejection classifier. In each 
round, we ran 3 million simulations.

In addition, it contains:
```python
simulate_11deg_R3_predictives_at_11deg
simulate_11deg_R3_predictives_at_27deg
```
which are posterior predictives of the 11deg posterior. Simulated at 11degree and 27 degree, respectively. At 11 degree, we have 17*6*10000 simulations. At 27 degree, as this is training data, we have 12*40*10000.

Lastly, it contains:
```python
simulate_27deg_R4_predictives_at_11deg
simulate_27deg_R4_predictives_at_27deg
```
They include the predictive samples from the posterior at 11deg+27deg. Each of them are around 1million simulations (17*6*10000).