search

Qengineering / YoloV5-ncnn-Jetson-Nano

YoloV5 for Jetson Nano
https://qengineering.eu/deep-learning-examples-on-raspberry-32-64-os.html
BSD 3-Clause "New" or "Revised" License
38 stars 8 forks source link

please can you try yolov5 nano #1

Open VYRION-Ai opened 2 years ago

VYRION-Ai commented 2 years ago

please can you try yolov5n and tell us the result fps

Qengineering commented 2 years ago

Can't find a suitable YoloV5n model for the ncnn framework. Due to many other projects I haven't time to convert and test one from another source.

VYRION-Ai commented 2 years ago

@Qengineering thank you very much I need your advice, i want ai kit to work with YOLO V5 model to work on real time project ( using camera) my budget is 200$ , what should I buy

Qengineering commented 2 years ago

You have two options. 1) Jetson Nano with 4 GByte RAM (not 2 GB) ± $ 120,= + $ 25,= for the camera. However, its very hard to get one due to the global chip shortage. Expected delivery in august this year (hopefully). You can try to get one second hand at e-bay. 2) OAK-D-Lite with three integrated camera's $ 200,= or OAK-1 for $ 150,= with one camera. Both have hardware acceleration on board. Not the Jetson CUDA, but the VPU Myriad processor. They are available: https://shop.luxonis.com/collections/all

Both options are state-of-the-art embedded systems. You must have at least some programming skills to work with it. The Jetson Nano has better support as it has been on the market for a while and has a good forum. OAK was founded in OpenCV, has been on the market for a relatively short time, but has many enthusiastic fans.

Example: https://blog.roboflow.com/deploy-luxonis-oak/