search

luxonis / depthai-experiments

Experimental projects we've done with DepthAI.
MIT License
799 stars 356 forks source link

Running YOLOv8 model on OAK D Pro Wide camera. I keep getting a not enough memory error. #483

Open ahmed-gaal opened 1 year ago

ahmed-gaal commented 1 year ago

i keep getting this error 2023-05-30 12:14:18] INFO [root.__init__:127] Setting IR laser dot projector brightness to 800mA [18443010D1A1840E00] [20.3.4.1.3] [7.950] [SpatialDetectionNetwork(2)] [error] Tried to allocate '38907200'B out of '5717503'B available. [18443010D1A1840E00] [20.3.4.1.3] [7.953] [SpatialDetectionNetwork(2)] [error] Neural network executor '0' out of '2' error: OUT_OF_MEMORY [18443010D1A1840E00] [20.3.4.1.3] [7.954] [SpatialDetectionNetwork(2)] [error] Tried to allocate '38907200'B out of '5717503'B available. [18443010D1A1840E00] [20.3.4.1.3] [7.959] [SpatialDetectionNetwork(2)] [error] Neural network executor '1' out of '2' error: OUT_OF_MEMORY [18443010D1A1840E00] [20.3.4.1.3] [7.959] [SpatialDetectionNetwork(2)] [error] Not enough memory to start a given neural network, needed at least: 38907200B [2023-05-30 12:14:28] INFO [root.__exit__:297] Closing OAK camera

Erol444 commented 1 year ago

Hi @ahmed-gaal , My guess is that you don't have enough memory:) Useful documentation in such case: https://docs.luxonis.com/projects/api/en/latest/tutorials/ram_usage/ But most of ram is likely consumed by your model. I'd go with small/tiny/nano arch of yolo.