MultiBotNav/transfer.py at main · ch0t0n/MultiBotNav · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
import os
import argparse
from datetime import datetime
import gymnasium as gym

from stable_baselines3.common.env_util import make_vec_env
from stable_baselines3.common.callbacks import LogEveryNTimesteps
from stable_baselines3.common.logger import configure

from src.utils import (
    load_model,
    read_wheeled_config,
    read_uav_json,
    parse_bool
)

if __name__ == '__main__':

    # ------------------------------------------------------------------
    # Ensure correct working directory
    # ------------------------------------------------------------------
    script_dir = os.path.dirname(os.path.realpath(__file__))
    os.chdir(script_dir)

    parser = argparse.ArgumentParser()

    parser.add_argument('--algorithm', type=str, required=True,
                        choices=['A2C', 'PPO', 'TRPO', 'TQC', 'ARS', 'CrossQ'])

    parser.add_argument('--robot_type', type=str,
                        choices=['uav', 'wheeled_robot'],
                        default='uav')

    parser.add_argument('--load_set', required=True, type=int)
    parser.add_argument('--train_set', required=True, type=int)

    parser.add_argument('--num_robots', type=int, default=3)

    parser.add_argument('--verbose', type=int, choices=[0, 1, 2], default=0)
    parser.add_argument('--steps', type=int, default=5_000_000)
    parser.add_argument('--num_envs', type=int, default=4)
    parser.add_argument('--seed', type=int, default=None)
    parser.add_argument('--log_steps', type=int, default=2000)
    parser.add_argument('--device', type=str,
                        choices=['cpu', 'cuda'],
                        default='cpu')

    args = parser.parse_args()
    print(args)

    if args.load_set == args.train_set:
        raise ValueError("load_set and train_set must be different.")

    # ------------------------------------------------------------------
    # Logging
    # ------------------------------------------------------------------
    log_dir = os.path.join(
        "logs",
        "transfer_logs",
        f"{args.robot_type}_{args.algorithm}_from{args.load_set}_to{args.train_set}"
    )
    os.makedirs(log_dir, exist_ok=True)

    logger = configure(
        os.path.join(log_dir, "tensorboard"),
        ["stdout", "log", "csv", "json", "tensorboard"]
    )

    # ------------------------------------------------------------------
    # Create training environment
    # ------------------------------------------------------------------
    if args.robot_type == "wheeled_robot":
        env_config = read_wheeled_config(
            f'exp_sets/wheeled/env{args.train_set}.ini'
        )

        vec_env = make_vec_env(
            'MultiWheeled-v0',
            env_kwargs={
                'render_mode': None,
                'env_params': env_config
            },
            n_envs=args.num_envs,
            seed=args.seed
        )

    else:
        env_json = read_uav_json(
            r'.\exp_sets\uav\icra_2026_cont_sets.json'
        )[f'set{args.train_set}']

        vec_env = make_vec_env(
            'MultiUAV-v0',
            env_kwargs={
                'render_mode': None,
                'field_info': env_json,
                'num_robots': args.num_robots
            },
            n_envs=args.num_envs,
            seed=args.seed
        )

    vec_env.action_space.seed(args.seed)

    # ------------------------------------------------------------------
    # Load model from training logs
    # ------------------------------------------------------------------
    trained_model_path = os.path.join(
        "logs",
        "training_default_logs",
        f"{args.robot_type}_{args.algorithm}_set{args.load_set}_seed{args.seed}_v0",
        "checkpoints",
        "trained_model.zip"
    )

    model = load_model(
        algorithm=args.algorithm,
        seed=args.seed,
        device=args.device,
        verbose=args.verbose,
        trained_model_path=trained_model_path
    )

    model.set_env(vec_env)
    model.set_logger(logger)

    # ------------------------------------------------------------------
    # Train (Transfer)
    # ------------------------------------------------------------------
    start_time = datetime.now()
    print(f"Transfer started on {start_time.ctime()}")

    callback = LogEveryNTimesteps(n_steps=args.log_steps)

    model.learn(
        total_timesteps=args.steps,
        callback=callback,
        log_interval=None,
        tb_log_name=f"{args.robot_type}_{args.algorithm}_transfer"
    )

    end_time = datetime.now()
    print(f"Transfer ended on {end_time.ctime()}")
    print(f"Transfer lasted {end_time - start_time}")

    # ------------------------------------------------------------------
    # Save model
    # ------------------------------------------------------------------
    save_dir = os.path.join(log_dir, "checkpoints")
    os.makedirs(save_dir, exist_ok=True)

    model.save(os.path.join(save_dir, "transfer_model.zip"))

    vec_env.close()