Notice
Recent Posts
Recent Comments
Link
| 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 |
Tags
- psuedo label
- deep learning on mobile
- chagpt
- cnn
- zero shot
- depth map
- 딥러닝 #deep learning #machine learning #computer vision #object detection
- sudo 권한
- convolution neural network
- semi supervised semantic segmentation
- GPT
- Deep Learning
- 파이썬
- 딥러닝
- 쿠버플로우
- domain adaptation
- mobile deep learning
- 고유값 분해
- segmentation
- semi supervised learning
- GPT 논문
- depth data
- deep learning mesh
- 수학
- GPT 1
- Computer Vision
- 3d
- 통계
- Python
- 머신러닝
Archives
- Today
- Total
튼튼한 나무
jupyter notebook에서 tensorboard 실시간으로 확인하기 본문
필자는 서버 굉장히 많이(아니 대부분) 사용하는데,
tensorboard로 scalar, image, audio, weight 등 다양한 값을 저장하고
서버에 연결하여 코딩도 하지만 juyter notebook으로 다양한 작업을 한다.
이런 경우에 tensorboard를 원격으로 접속가능하게끔 띄워도 되지만 뭔가 굉장히 귀찮다..
이걸 또 안하려니 매번 tensorboard file을 로컬에 다운받아서 tensorboard를 로컬에 띄워야 한다. (매우 귀찮다..)
그래서 jupyter notebook에서 tensorboard를 실시간으로 띄워보고자 한다.
먼저 port를 뚫어준다. (아래 예시는 6006~6020 port open)
1) ubuntu server
sudo ufw allow 6006:6020/tcp
2) Docker로 Jupyter Notebook 서버를 사용 한다면, Docker 에서 외부 -> 내부 컨테이너 포워딩 설정.
docker run --runtime=nvidia --rm -itd -p 8888:8888 -p 6006-6020:6006-6020 -v (local volumne):(도커컨테이너에 연결할 volumne) jupyter notebook
Jupyter example


'computing' 카테고리의 다른 글
| 모바일 배포를 위해 pytorch 및 tensorflow 모델 변환이 필요한 이유 완전 정복하기 (feat. onnx or tflite) (0) | 2023.07.11 |
|---|---|
| 파이썬 코드로 GPU memory print 하기 (0) | 2023.03.01 |
| 파이썬으로 sudo권한 shell 명령어 실행하기 (0) | 2023.01.02 |
| docker에서 한국 시간으로 날짜 설정하기 (0) | 2022.03.22 |
| DistributedDataParallel(DDP)를 이용한 pytorch model save error (0) | 2022.03.21 |
Comments