Sign inSign up

juhyung1021/vscode-torch

By juhyung1021

•Updated over 1 year ago

Has VsCodeServer, Torch, Miniconda, jupyter notebook

Image
1

274

juhyung1021/vscode-torch repository overview

⁠Detail

https://github.com/SagiK-Repository/AI_Deep_Learn/releases/tag/localpytorchdocker⁠


⁠Fast Run

  • 아래 명령어 실행만 하면 바로 사용할 수 있습니다.
# cuda 12.2.2
# 웬만하면 12.2.2로 돌아갑니다.
docker run -it --gpus all --name vscode-torch -p 18094:8080 -d juhyung1021/vscode-torch:cuda12.2.2

# cuda 12.6.3
docker run -it --gpus all --name vscode-torch -p 18094:8080 -d juhyung1021/vscode-torch:cuda12.6.3

⁠Download

⁠Error Cases

⁠[Error Case 01] Cuda Error 발생하는 경우
docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy'
nvidia-container-cli: requirement error: unsatisfied condition: cuda>=12.6, please update your driver to a newer version, or use an earlier cuda container: unknown.
  • GPU CUDA Version을 확인 후 dockerfile 이미지의 tag를 검토
  • 지원하는 버전이 없으면 직접 dockerfile 배포
⁠[Error Case 02] Cuda 못찾는 경우
torch.cuda.is_available() == false
torch.cuda.device_count() > 0인 경우

Tag summary

Content type

Image

Digest

sha256:8bb15d3cf…

Size

10.6 GB

Last updated

over 1 year ago

docker pull juhyung1021/vscode-torch:cuda12.6.3