[Unity] ML-Agent 버전 맞추기 (Window)
이거 세팅해두는게 사실 젤 눈물납니다.
그래픽 카드 있는데도 CPU 버전으로 하시는분들 계시죠? 그러지맙시다..
CPU 터질라해요!!
버전 정리
| GPU | RTX3060 |
| Unity | 2022.3.17f1 |
| ML-Agents | 릴리스 19 |
| Conda | |
| Python | 3.8 |
| Cuda | 11.3 |
| Torch | 1.10.0+cu113 |
| onnx | 1.16.1 |
| numpy | 1.19.5 |
| protobuf | 3.20.2 |
크게 필요했던거 정리해뒀습니다.
!! Cuda는 본인 GPU에 맞는거 설치하셔야해요..
Unity
2022.3.17f1 LTS 버전
>> Archive
다운로드 아카이브
다운로드 아카이브
unity.com
ML-Agents
릴리스 19 버전
>> GitHub
GitHub - Unity-Technologies/ml-agents: The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enab
The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enables games and simulations to serve as environments for training intelligent agents using deep reinforcement ...
github.com
Conda
가상환경 구성 // 여기서 파이썬 3.8버전 설치
>> 공식 홈페이지
Download Now | Anaconda
Anaconda is the birthplace of Python data science. We are a movement of data scientists, data-driven enterprises, and open source communities.
www.anaconda.com
conda create -n test python=3.8
conda activate test
CUDA
11.3
>> 공식 홈페이지
CUDA Toolkit 11.3 Downloads
Get CUDA Toolkit 11.3 for Linux and Windows.
developer.nvidia.com
pip install torch==1.10.0+cu113 torchvision==0.11.1+cu113 torchaudio===0.10.0+cu113 -f https://download.pytorch.org/whl/torch_stable.html
전체 버전
(test) C:\Users\user\ml-agents-release_19>conda list
# packages in environment at C:\Users\user\anaconda3\envs\test:
#
# Name Version Build Channel
absl-py 2.1.0 pypi_0 pypi
attrs 23.2.0 pypi_0 pypi
blas 1.0 mkl
brotli-python 1.0.9 py38hd77b12b_8
ca-certificates 2024.7.2 haa95532_0
cachetools 5.4.0 pypi_0 pypi
cattrs 1.5.0 pypi_0 pypi
certifi 2024.7.4 py38haa95532_0
charset-normalizer 3.3.2 pyhd3eb1b0_0
cloudpickle 3.0.0 pypi_0 pypi
cuda-cccl 12.4.127 0 nvidia
cuda-cudart 12.4.127 0 nvidia
cuda-cudart-dev 12.4.127 0 nvidia
cuda-cupti 12.4.127 0 nvidia
cuda-libraries 12.4.0 0 nvidia
cuda-libraries-dev 12.4.0 0 nvidia
cuda-nvrtc 12.4.127 0 nvidia
cuda-nvrtc-dev 12.4.127 0 nvidia
cuda-nvtx 12.4.127 0 nvidia
cuda-opencl 12.4.127 0 nvidia
cuda-opencl-dev 12.4.127 0 nvidia
cuda-profiler-api 12.4.127 0 nvidia
cuda-runtime 12.4.0 0 nvidia
cudatoolkit 11.3.1 h59b6b97_2
filelock 3.15.4 pypi_0 pypi
freetype 2.12.1 ha860e81_0
gmpy2 2.1.2 py38h7f96b67_0
google-auth 2.32.0 pypi_0 pypi
google-auth-oauthlib 1.0.0 pypi_0 pypi
grpcio 1.65.1 pypi_0 pypi
gym 0.26.2 pypi_0 pypi
gym-notices 0.0.8 pypi_0 pypi
h5py 3.11.0 pypi_0 pypi
icc_rt 2022.1.0 h6049295_2
idna 3.7 py38haa95532_0
importlib-metadata 8.2.0 pypi_0 pypi
intel-openmp 2023.1.0 h59b6b97_46320
jinja2 3.1.4 py38haa95532_0
jpeg 9e h827c3e9_2
lcms2 2.12 h83e58a3_0
lerc 3.0 hd77b12b_0
libcublas 12.4.2.65 0 nvidia
libcublas-dev 12.4.2.65 0 nvidia
libcufft 11.2.0.44 0 nvidia
libcufft-dev 11.2.0.44 0 nvidia
libcurand 10.3.5.147 0 nvidia
libcurand-dev 10.3.5.147 0 nvidia
libcusolver 11.6.0.99 0 nvidia
libcusolver-dev 11.6.0.99 0 nvidia
libcusparse 12.3.0.142 0 nvidia
libcusparse-dev 12.3.0.142 0 nvidia
libdeflate 1.17 h2bbff1b_1
libffi 3.4.4 hd77b12b_1
libjpeg-turbo 2.0.0 h196d8e1_0
libnpp 12.2.5.2 0 nvidia
libnpp-dev 12.2.5.2 0 nvidia
libnvfatbin 12.4.127 0 nvidia
libnvfatbin-dev 12.4.127 0 nvidia
libnvjitlink 12.4.99 0 nvidia
libnvjitlink-dev 12.4.99 0 nvidia
libnvjpeg 12.3.1.89 0 nvidia
libnvjpeg-dev 12.3.1.89 0 nvidia
libpng 1.6.39 h8cc25b3_0
libprotobuf 3.20.1 h23ce68f_0
libtiff 4.5.1 hd77b12b_0
libuv 1.48.0 h827c3e9_0
libwebp-base 1.3.2 h2bbff1b_0
lz4-c 1.9.4 h2bbff1b_1
markdown 3.6 pypi_0 pypi
markupsafe 2.1.5 pypi_0 pypi
mkl 2023.1.0 h6b88ed4_46358
mkl-service 2.4.0 py38h2bbff1b_1
mkl_fft 1.3.8 py38h2bbff1b_0
mkl_random 1.2.4 py38h59b6b97_0
mlagents 0.30.0 pypi_0 pypi
mlagents-envs 0.30.0 pypi_0 pypi
mpc 1.1.0 h7edee0f_1
mpfr 4.0.2 h62dcd97_1
mpir 3.0.0 hec2e145_1
mpmath 1.3.0 py38haa95532_0
networkx 3.1 py38haa95532_0
numpy 1.19.5 pypi_0 pypi
oauthlib 3.2.2 pypi_0 pypi
onnx 1.16.1 pypi_0 pypi
openjpeg 2.4.0 h4afccc4_2
openssl 3.0.14 h827c3e9_0
pettingzoo 1.15.0 pypi_0 pypi
pillow 10.4.0 py38h827c3e9_0
pip 24.0 py38haa95532_0
protobuf 3.20.2 pypi_0 pypi
pyasn1 0.6.0 pypi_0 pypi
pyasn1-modules 0.4.0 pypi_0 pypi
pypiwin32 223 pypi_0 pypi
pysocks 1.7.1 py38haa95532_0
python 3.8.19 h1aa4202_0
pytorch-cuda 12.4 h3fd98bf_6 pytorch
pytorch-mutex 1.0 cuda pytorch
pywin32 306 pypi_0 pypi
pyyaml 6.0.1 py38h2bbff1b_0
requests 2.32.3 py38haa95532_0
requests-oauthlib 2.0.0 pypi_0 pypi
rsa 4.9 pypi_0 pypi
setuptools 69.5.1 py38haa95532_0
six 1.16.0 pypi_0 pypi
sqlite 3.45.3 h2bbff1b_0
sympy 1.12 py38haa95532_0
tbb 2021.8.0 h59b6b97_0
tensorboard 2.14.0 pypi_0 pypi
tensorboard-data-server 0.7.2 pypi_0 pypi
torch 1.10.0+cu113 pypi_0 pypi
torchaudio 0.10.0+cu113 pypi_0 pypi
torchvision 0.11.1+cu113 pypi_0 pypi
typing_extensions 4.11.0 py38haa95532_0
urllib3 2.2.2 py38haa95532_0
vc 14.2 h2eaa2aa_4
vs2015_runtime 14.29.30133 h43f2093_4
werkzeug 3.0.3 pypi_0 pypi
wheel 0.43.0 py38haa95532_0
win_inet_pton 1.1.0 py38haa95532_0
xz 5.4.6 h8cc25b3_1
yaml 0.2.5 he774522_0
zipp 3.19.2 pypi_0 pypi
zlib 1.2.13 h8cc25b3_1
zstd 1.5.5 hd43e919_2
GPU 연결 확인
CUDA가 설치가 잘됐고 사용가능한지 보는 파이썬 코드도 올려드리겠습니다.
import torch
if torch.cuda.is_available():
device = torch.device("cuda")
print("GPU 사용:", torch.cuda.get_device_name(0))
else:
device = torch.device("cpu")
print("GPU 사용 불가능")
이런식으로 사용하시면 됩니다.
참고 블로그
Unity ML-Agent 설치하기
이번 포스팅은 가상공간에서 강화학습을 이용해 인공지능을 학습할 수 있는 환경을 제공하는 Unity ML-Agent를 설치하는 과정입니다. 이 포스트는 아래 내용을 참고했습니다. github: https://github.com/Un
with-rl.tistory.com
>> ml-agent 설정
ml-agent 설정
WSL2에서 mlagent 설치시 이런 저런 dependency를 강하게 받는다0.30.0 버전 설치시 아래의 요구사항이 있다https://unity-technologies.github.io/ml-agents/Installation/python 3.8.13 ~
velog.io