hotfix/setup_for_window (#98)

* hotfix/setup_for_window

* .

README.md

@@ -55,18 +55,9 @@ https://www.anaconda.com/
 
 **Windows 사용자**
 
-git bash에서 다음 명령어를 실행시켜 conda 명령어가 동작하도록 합니다.
-
 ```bash
-$ echo ". /c/Users/{유저이름}/anaconda3/etc/profile.d/conda.sh" >> ~/.profile
-```
-
-git bash를 재실행하고 다음 명령어를 차례대로 실행시켜 주세요.
-
-```bash
-$ conda activate meta
-
-(meta) $ sh ./scripts/window-init.sh
+# 사용자
+(meta) $ "./scripts/window-init.bat"
 ```
 
 ### 4. 모델 학습 및 결과 확인

requirements.txt

@@ -1,11 +1,11 @@
 autopep8==1.5.0
 GPUtil==1.4.0
-gym>=0.24.1
+gym==0.24.1
 imageio>=2.1.2
 jupyter==1.0.0
 jupyter-contrib-nbextensions==0.5.1
 jupyter-nbextensions-configurator==0.4.1
 matplotlib>=3.5.2
-mujoco>=2.2.0
+mujoco==2.2.0
 torchmeta>=1.8.0
 tqdm>=4.62.3

scripts/window-init.sh → scripts/window-init.bat

@@ -1,5 +1,3 @@
-#!/bin/bash
-
 pip install -e .
 pip install -r requirements.txt
 python ./scripts/download-torch.py

src/meta_rl/envs/__init__.py

@@ -1,10 +1,11 @@
 import importlib
 import os
+from typing import Callable
 
 ENVS = {}
 
 
-def register_env(name: str) -> function:
+def register_env(name: str) -> Callable:
  def register_env_fn(filename: str) -> str:
  if name in ENVS:
  raise ValueError(f"Cannot register duplicate env {name}")

src/meta_rl/maml/algorithm/networks.py

@@ -63,8 +63,8 @@ def __init__(
 
  self.log_std = torch.Tensor([init_std]).log()
  self.log_std = torch.nn.Parameter(self.log_std)
- self.min_log_std = torch.Tensor([min_std]).log().item() if not min_std else None
- self.max_log_std = torch.Tensor([max_std]).log().item() if not max_std else None
+ self.min_log_std = torch.Tensor([min_std]).log().item() if min_std is not None else None
+ self.max_log_std = torch.Tensor([max_std]).log().item() if max_std is not None else None
 
  self.is_deterministic = is_deterministic
 

src/meta_rl/maml/maml_trainer.py

@@ -12,13 +12,14 @@
 
 if __name__ == "__main__":
  # 실험 환경 설정에 대한 하이퍼파라미터들 불러오기
- with open(os.path.join("configs", "experiment_config.yaml"), "r") as file:
+ with open(os.path.join("configs", "experiment_config.yaml"), "r", encoding="utf-8") as file:
  experiment_config: Dict[str, Any] = yaml.load(file, Loader=yaml.FullLoader)
 
  # 목표 보상 설정에 대한 하이퍼파라미터들 불러오기
  with open(
  os.path.join("configs", experiment_config["env_name"] + "_target_config.yaml"),
  "r",
+ encoding="utf-8",
  ) as file:
  env_target_config: Dict[str, Any] = yaml.load(file, Loader=yaml.FullLoader)
 

src/meta_rl/pearl/pearl_trainer.py

@@ -12,13 +12,14 @@
 
 if __name__ == "__main__":
  # 실험 환경 설정에 대한 하이퍼파라미터들 불러오기
- with open(os.path.join("configs", "experiment_config.yaml"), "r") as file:
+ with open(os.path.join("configs", "experiment_config.yaml"), "r", encoding="utf-8") as file:
  experiment_config: Dict[str, Any] = yaml.load(file, Loader=yaml.FullLoader)
 
  # 목표 보상 설정에 대한 하이퍼파라미터들 불러오기
  with open(
  os.path.join("configs", experiment_config["env_name"] + "_target_config.yaml"),
  "r",
+ encoding="utf-8",
  ) as file:
  env_target_config: Dict[str, Any] = yaml.load(file, Loader=yaml.FullLoader)
 

src/meta_rl/rl2/rl2_trainer.py

@@ -12,13 +12,14 @@
 
 if __name__ == "__main__":
  # 실험 환경 설정에 대한 하이퍼파라미터들 불러오기
- with open(os.path.join("configs", "experiment_config.yaml"), "r") as file:
+ with open(os.path.join("configs", "experiment_config.yaml"), "r", encoding="utf-8") as file:
  experiment_config: Dict[str, Any] = yaml.load(file, Loader=yaml.FullLoader)
 
  # 목표 보상 설정에 대한 하이퍼파라미터들 불러오기
  with open(
  os.path.join("configs", experiment_config["env_name"] + "_target_config.yaml"),
  "r",
+ encoding="utf-8",
  ) as file:
  env_target_config: Dict[str, Any] = yaml.load(file, Loader=yaml.FullLoader)