eval_rl.py

import os
USED_DEVICES = "1"  # if your want to use CPU in a server with GPU, change "0" to "-1"
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = USED_DEVICES
os.environ["CUDA_LAUNCH_BLOCKING"] = "1"

import torch
import numpy as np
import random

seed = 0  # use the fixed seed for the full program

# must use
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed)  # if you are using multi-GPU.

# optional use
# torch.set_deterministic(True)
# torch.backends.cudnn.enabled = False 
# torch.backends.cudnn.benchmark = False
# os.environ['CUBLAS_WORKSPACE_CONFIG'] = ':4096:8'
os.environ['PYTHONHASHSEED'] = str(seed)

import alphastarmini

from alphastarmini.core.arch import entity_encoder
from alphastarmini.core.arch import scalar_encoder
from alphastarmini.core.arch import spatial_encoder
from alphastarmini.core.arch import arch_model
from alphastarmini.core.arch import action_type_head
from alphastarmini.core.arch import selected_units_head
from alphastarmini.core.arch import target_unit_head
from alphastarmini.core.arch import delay_head
from alphastarmini.core.arch import queue_head
from alphastarmini.core.arch import location_head
from alphastarmini.core.arch import agent
from alphastarmini.core.arch import baseline

from alphastarmini.core.sl import load_pickle

from alphastarmini.core.rl import action
from alphastarmini.core.rl import env_utils
from alphastarmini.core.rl import actor
from alphastarmini.core.rl import against_computer
from alphastarmini.core.rl import pseudo_reward

import param as P

if __name__ == '__main__':
    # if we don't add this line, it may cause running time error while in Windows
    # torch.multiprocessing.freeze_support()

    print("run init")

    # ------------------------

    # 3. we use RL environment to multi-process (thread) evaluate SL model
    from alphastarmini.core.rl import mp_rl_eval_rl
    mp_rl_eval_rl.test(on_server=P.on_server)

    print('run over')