Move models into separate directories

[damonbayer]

Output structure looks like

├── data
│   ├── data_for_model_fit.json
│   ├── data.tsv
│   ├── epiweekly_data.tsv
│   ├── epiweekly_eval_data.tsv
│   └── eval_data.tsv
├── Disease_forecast_plot_log.pdf
├── Disease_forecast_plot.pdf
├── epiweekly_score_table.rds
├── Other_forecast_plot_log.pdf
├── Other_forecast_plot.pdf
├── priors.py
├── prop_disease_ed_visits_forecast_plot_log.pdf
├── prop_disease_ed_visits_forecast_plot.pdf
├── pyrenew_e
│   ├── combined_training_eval_data.parquet
│   ├── epiweekly_forecast_samples.parquet
│   ├── forecast_ci.parquet
│   ├── forecast_samples.parquet
│   ├── inference_data.csv
│   ├── inference_data.nc
│   ├── mcmc_tidy
│   │   ├── pyrenew_log_likelihood.parquet
│   │   ├── pyrenew_posterior.parquet
│   │   ├── pyrenew_posterior_predictive.parquet
│   │   └── pyrenew_sample_stats.parquet
│   └── posterior_samples.pickle
├── score_table.rds
└── timeseries_e
    ├── baseline_cdc_count_ed_visits_forecast.parquet
    ├── baseline_cdc_prop_ed_visits_forecast.parquet
    ├── baseline_ts_count_ed_visits_forecast.parquet
    ├── baseline_ts_prop_ed_visits_forecast.parquet
    ├── epiweekly_baseline_cdc_count_ed_visits_forecast.parquet
    ├── epiweekly_baseline_cdc_prop_ed_visits_forecast.parquet
    ├── epiweekly_baseline_ts_count_ed_visits_forecast.parquet
    ├── epiweekly_baseline_ts_prop_ed_visits_forecast.parquet
    ├── epiweekly_other_ed_visits_forecast.parquet
    └── other_ed_visits_forecast.parquet

[damonbayer]

@dylanhmorris Ready for your review

[dylanhmorris]

LGTM