vuong02011996 / ai-engineer

ai-engineer

• Maths-basic with fine-turning model.
• Convert model to tensorRT.
• Docker - build docker file, service model.
• Opencv - some function.(watermark, preprocessing for face recognition)
• Pandas
• BBGT + clover ok

Note

• Pytorch much match with cuda version
• link

conda install pytorch==1.7.1 torchvision==0.8.2 torchaudio==0.7.2 cudatoolkit=10.1 -c pytorch

Link ref

• python_trending

Ref a Thanh

System to maintain camera AI

• Media pipe - https://github.com/google/mediapipe?tab=readme-ov-file - Google
• dlstreamer - https://dlstreamer.github.io/ - Intel
• Deepstreamer - https://github.com/NVIDIA-AI-IOT/deepstream_python_apps - Nvidia
• Gstreamer - FFMPEG
• OpenCV graph
• NVIDIA DRIVE End-to-End Solutions for Autonomous Vehicles - https://developer.nvidia.com/drive
• A General-purpose Parallel and Heterogeneous Task Programming System - https://github.com/taskflow/taskflow
• trild-vietnam - Isarel
• NVIDIA Metropolis Microservices - https://developer.nvidia.com/metropolis-microservices
• Nvidia City: https://github.com/NVIDIAAICITYCHALLENGE/2023AICITY_Code_From_Top_Teams?tab=readme-ov-file

Celery

• Van de hien tai: Khi goi api toi model xu li, api sau phai doi api truoc xu li xong moi lam viec duoc.
• Cach su ly bat dong bo khi goi API.

Q/A a Thanh

• How to write document: api
• CI/CD gitlab
  • setup to create pull request
  • aprrove
  • create issue