| import os |
| import random |
| import numpy as np |
| import torch as th |
|
|
|
|
| def set_seed(seed): |
| random.seed(seed) |
| np.random.seed(seed) |
| th.manual_seed(seed) |
| th.cuda.manual_seed(seed) |
| th.backends.cudnn.deterministic = True |
| th.backends.cudnn.benchmark = False |
| os.environ['PYTHONHASHSEED'] = str(seed) |
|
|
|
|
| def save_param_info(folder_path, algo_name, params): |
| |
| reward_types = params.reward_types |
| use_adrs = params.use_adrs |
| |
| ep_step = params.episode_step |
| hybrid_eta = params.hybrid_eta |
| adrs_mu = params.adrs_mu |
| adrs_update = params.adrs_update |
| gamma = params.gamma |
| total_timesteps = params.total_timesteps |
| exploration_initial_eps = params.exploration_initial_eps, |
| exploration_fraction = params.exploration_fraction, |
| exploration_final_eps = params.exploration_final_eps, |
| map_id = params.map_id |
| env_name = params.env_name |
| lr_start = params.learning_rate_start |
| lr_end = params.learning_rate_end |
| lr_fraction = params.learning_fraction |
|
|
| model_params = params.model_params(algo_name) |
|
|
| with open(folder_path + "/" + "model param info.txt", "w") as f: |
| f.write("env name: " + str(env_name) + "\n") |
| f.write("map id: " + str(map_id) + "\n") |
| f.write("algorithm name: " + algo_name + "\n") |
| f.write("reward types: " + str(reward_types) + "\n") |
| f.write("use adrs: " + str(use_adrs) + "\n") |
| f.write("episode step: " + str(ep_step) + "\n") |
| f.write("adrs mu: " + str(adrs_mu) + "\n") |
| f.write("hybrid eta: " + str(hybrid_eta) + "\n") |
| f.write("adrs update: " + str(adrs_update) + "\n") |
| f.write("gamma: " + str(gamma) + "\n") |
| f.write("total_timesteps: " + str(total_timesteps) + "\n") |
| f.write("exploration_final_eps: " + str(exploration_final_eps) + "\n") |
| f.write("exploration_initial_eps: " + str(exploration_initial_eps) + "\n") |
| f.write("exploration_fraction: " + str(exploration_fraction) + "\n") |
| f.write("learning rate start: " + str(lr_start) + "\n") |
| f.write("learning_rate_end: " + str(lr_end) + "\n") |
| f.write("learning_rate_fraction: " + str(lr_fraction) + "\n") |
| f.write("seed: " + str(params.seed) + "\n") |
| f.write("noise_level: " + str(params.noise_level) + "\n") |
| |
| for item in model_params.items(): |
| f.write(str(item)) |
|
|
|
|
| def save_model_param_info(model, folder_path, algo_name, params): |
| if not os.path.exists(folder_path): |
| os.makedirs(folder_path) |
| save_path = folder_path + "/" + algo_name |
| model.save(save_path) |
|
|
| save_param_info(folder_path, algo_name, params) |
|
|
