0. Configure environment and DB.

   • Copy config.py.template to config.py and fill in missing entries to reflect your system.
   • Do the same for db/credentials.py
   • run python setup.py install

1. Create an experiment.

   • See experiments/nist_baseline.py for an example experiment definition.
   • See models/seung_unet.py for an example model specification (included in the experiment definition).
   • See datasets/cluttered_nist_baseline.py for an example dataset class (included in the experiment definition).
   • Initialize the DB and load an experiment: python build_experiments.py --experiment=nist_baseline --initialize
     • I manually access the db with psql cluttered_nist -h 127.0.0.1 -d cluttered_nist.

2. Run an experiment.

   • A single job from the DB: CUDA_VISIBLE_DEVICES=0 python run_job.py
   • A single job without the DB: CUDA_VISIBLE_DEVICES=0 python run_job.py --no_db --experiment=nist_baseline --model=seung_unet --train=cluttered_nist_baseline --test=cluttered_nist_baseline
   • A local worker that continues until the DB is exhausted: bash start_worker.sh
   • Fill the p-nodes with workers running in Dockers: bash docker_workers.sh

3. Manually access the DB - psql cluttered_nist_p7 -h 127.0.0.1 -d cluttered_nist_p7

4. Run and kill docker jobs - Run docker job bash docker_workers.sh - Kill docker jobs python utils/docker_kill.py bash - Get docker pids docker ps - Get docker job stdout `docker logs

---

Extract BSDS tuning data

• cd /media/data_cifs/projects/prj_neural_circuits/bsds_ibm
• CUDA_VISIBLE_DEVICES=2 python run_job.py --experiment=BSDS500_test --model=BSDS_vgg_cheap_deepest_final_simple --no_db --ckpt=/media/data_cifs/cluttered_nist_experiments/checkpoints/BSDS_vgg_cheap_deepest_final_simple_BSDS500_combos_100_hed_flips_thresh_2_2019_07_08_20_42_03_800563/model_152400.ckpt-152400 --test --out_dir=bsds_landscape --train=BSDS500_test_landscape --val=BSDS500_test_landscape --placeholders
• Transfer to the mac and run bsds_tuning.py