zjowowen's picture
init space
079c32c
from typing import Union, Optional, List, Any, Callable, Tuple
import torch
from ding.config import compile_config, read_config
from ding.envs import get_vec_env_setting
from ding.policy import create_policy
from ding.utils import set_pkg_seed
from ding.torch_utils import to_tensor, to_ndarray, tensor_to_list
def eval(
input_cfg: Union[str, Tuple[dict, dict]],
seed: int = 0,
model: Optional[torch.nn.Module] = None,
state_dict: Optional[dict] = None,
replay_path: Optional[str] = './video',
) -> float:
r"""
Overview:
The evaluation entry for NGU policy.
Arguments:
- input_cfg (:obj:`Union[str, Tuple[dict, dict]]`): Config in dict type. \
``str`` type means config file path. \
``Tuple[dict, dict]`` type means [user_config, create_cfg].
- seed (:obj:`int`): Random seed.
- env_setting (:obj:`Optional[List[Any]]`): A list with 3 elements: \
``BaseEnv`` subclass, collector env config, and evaluator env config.
- model (:obj:`Optional[torch.nn.Module]`): Instance of torch.nn.Module.
- state_dict (:obj:`Optional[dict]`): The state_dict of policy or model.
"""
if isinstance(input_cfg, str):
cfg, create_cfg = read_config(input_cfg)
else:
cfg, create_cfg = input_cfg
create_cfg.policy.type += '_command'
cfg = compile_config(cfg, auto=True, create_cfg=create_cfg)
env_fn, _, evaluator_env_cfg = get_vec_env_setting(cfg.env)
env = env_fn(evaluator_env_cfg[0])
env.seed(seed, dynamic_seed=False)
set_pkg_seed(seed, use_cuda=cfg.policy.cuda)
policy = create_policy(cfg.policy, model=model, enable_field=['eval']).eval_mode
if state_dict is None:
state_dict = torch.load(cfg.learner.load_path, map_location='cpu')
policy.load_state_dict(state_dict)
env.enable_save_replay(replay_path=replay_path)
obs = env.reset()
obs = {0: obs}
episode_return = 0.
beta_index = {i: 0 for i in range(1)}
beta_index = to_tensor(beta_index, dtype=torch.int64)
prev_action = {i: torch.tensor(-1) for i in range(1)}
prev_reward_e = {i: to_tensor(0, dtype=torch.float32) for i in range(1)}
while True:
# TODO(pu): r_i, reward embedding
policy_output = policy.forward(beta_index, obs, prev_action, prev_reward_e)
actions = {i: a['action'] for i, a in policy_output.items()}
actions = to_ndarray(actions)
action = policy_output[0]['action']
action = to_ndarray(action)
timestep = env.step(action)
# print(action)
# print(timestep.reward)
timesteps = {0: timestep}
timesteps = to_tensor(timesteps, dtype=torch.float32)
prev_reward_e = {env_id: timestep.reward for env_id, timestep in timesteps.items()}
prev_reward_e = to_ndarray(prev_reward_e)
prev_action = actions
timestep = timesteps[0]
# print(timestep.info)
episode_return += timestep.reward
obs = timestep.obs
obs = {0: obs}
if timestep.done:
print(timestep.info)
break
print('Eval is over! The performance of your RL policy is {}'.format(episode_return))
if __name__ == "__main__":
# Users should add their own model path here. Model path should lead to a model.
# Absolute path is recommended.
# In DI-engine, it is ``exp_name/ckpt/ckpt_best.pth.tar``.
model_path = './debug_minigrid_doorkey_ngu_ul298_er01_n32_rbs3e4_fixepseval/ckpt/ckpt_best.pth.tar',
# model_path = 'model_path_placeholder',
cfg = '../config/minigrid_ngu_config.py'
state_dict = torch.load(model_path, map_location='cpu')
for i in range(0, 10):
eval(cfg, seed=i, state_dict=state_dict, replay_path='./video')