evo_train_cpu.sh

#!/bin/bash
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=12
##SBATCH --gres=gpu:1
#SBATCH --time=24:00:00
#SBATCH --mem=50GB
#SBATCH --job-name=nmmo_eval_2
#SBATCH --mail-type=BEGIN,END
#SBATCH --mail-user=sam.earle@nyu.edu
#SBATCH --output=nmmo_eval_2_%j.out

cd /scratch/se2161/neural-mmo || exit

#conda init bash
source activate
conda activate nmmo

export TUNE_RESULT_DIR='./ray_results/'
python Forge.py evaluate --config TreeOrerock --MAP fit-Differential_skills-ALL_gene-Simplex_algo-MAP-Elites_0 --MODELS ['fit-Lifespans_skills-NONE_gene-Baseline_algo-MAP-Elites_0','fit-Lifespans_skills-NONE_gene-Simplex_algo-MAP-Elites_0','fit-Differential_skills-ALL_gene-Simplex_algo-MAP-Elites_0'] --NPOLICIES 3 --NPOP 3 --PAIRED False --INFER_IDXS "[(32, 19, 0), (24, 14, 0), (22, 20, 0), (42, 19, 0)]"  --EVALUATION_HORIZON 500 --N_EVAL 5 --N_EVAL_MAPS 4 --NEW_EVAL --SKILLS "['constitution', 'fishing', 'hunting', 'range', 'mage', 'melee', 'defense', 'woodcutting', 'mining', 'exploration',]" --NENT 16 --FITNESS_METRIC Lifespans  --EVO_DIR 0 --COMPETITIVE_EVAL True

# need a blank space say the elders