
1. 터미널에서 도커 컨테이너 실행 (-v 옵션으로 공유 디렉토리를 설정해 놓으면 로컬폴더에 연결 가능하다!) https://nuggy875.tistory.com/49 Docker run 옵션 정리 docker run 명령어 $ docker run [Options] [Image] [Command] [Options] -it -i : interactive, -t ; tty 이 옵션을 줘야 컨테이너 안에서 터미널 실행 가능 -p [호스트의 포트:컨테이너 포트] -p 8888:8888 도커.. nuggy875.tistory.com 2. VSCode Extention인 'Remote Development' 설치 3. VSCode에서 Ctrl + Shift + p 누르고 'Remote-Containers: Atta..
docker run 명령어 $ docker run [Options] [Image] [Command] [Options] -it -i : interactive, -t ; tty 이 옵션을 줘야 컨테이너 안에서 터미널 실행 가능 -p [호스트의 포트:컨테이너 포트] -p 8888:8888 도커 호스트의 포트 : 컨테이너 내부 포트 Tensorboard등의 실행에 염두해 포트를 여러개 열어둘 수도 있음 -p 8888:8888 -p 6006-6015:6006-6015 --name [컨테이너 이름] --name pt -v [Host Path : Container Path] -v ${HOME}/dev:/data 보여질 호스트 폴더 : 컨테이너의 폴더 --gpus all Host의 Nvidia드라이버가 컨테이너에도 적..
ubuntu 18.04 RTX 2080 cuda 9.2 cudnn 7.6.4 pytorch 1.3.0 nvcc fatal unsupported gpu architecture 'compute_75' 에러 발생... 아래와 같이 nvcc compiler가 더 낮은 버전의 compute capability를 사용하도록 임의로 조정해주면 해결 된다. export TORCH_CUDA_ARCH_LIST="7.0" 출처 : https://github.com/torch/torch7/issues/1190
Cuda Remove sudo apt-get --purge remove 'cuda*' sudo apt-get autoremove --purge 'cuda*' /usr/local/ 폴더에 cuda 폴더가 남아있으면 지워주자 cuda든 cuda_9.x 든 뭐든 몽땅! CUDA 설치 https://developer.nvidia.com/cuda-toolkit-archive 깔고자 하는 버전을 선택하여 .run file 다운로드 빨간 글씨는 버전에 맞게 수정하여 실행 cd ~/Downloads/ sudo mv cuda_10.0*.* ~/ sudo chmod +x cuda_10.0*.* ./cuda_10.0.176_384.81_linux.run --override EULA License에서는 그래픽드라이버는 설치하..
- Total
- Today
- Yesterday
- Neural Radiance Field
- nginx
- GaussianSplatting
- ubuntu
- numpy
- MacOS
- Object Detection
- git
- SSH
- nvidia
- CUDA
- 우분투
- Deep Learning
- Anaconda
- Docker
- java
- pytorch
- Novel View Synthesis
- 2-stage Detector
- Computer Vision
- error
- 3Dvision
- Macbook
- MySQL
- Python
- vscode
- Machine Learning
- GPU
- Android
- nerf
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
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 |