CodeProject AI Docker Container not using CUDA GPU
-
Installed a Tesla P4 GPU on my Unraid server. The GPU is working for my Plex server running as docker container. Installed CodeProject.AI GPU docker container, the GPU is detected, however it is not being used for detections. System Info: Server version: 2.2.4-Beta Operating System: Linux (Linux 5.19.17-Unraid #2 SMP PREEMPT_DYNAMIC Wed Nov 2 11:54:15 PDT 2022) GPU: Tesla P4 (8 GiB) (NVIDIA) Driver: 535.54.03 CUDA: 12.2 (max supported: 12.2) Compute: 6.1 .NET framework: .NET 7.0.11 Video adapter info: System GPU info: GPU 3D Usage 0% GPU RAM Usage 76 MiB Global Environment variables: CPAI_APPROOTPATH = CPAI_PORT = 32168 Server Logs: 22:09:18:Operating System: Linux (Linux 5.19.17-Unraid #2 SMP PREEMPT_DYNAMIC Wed Nov 2 11:54:15 PDT 2022) 22:09:18:.NET framework: .NET 7.0.11 22:09:18:App DataDir: /etc/codeproject/ai 22:09:18:Video adapter info: 22:09:18:STARTING CODEPROJECT.AI SERVER 22:09:18:RUNTIMES_PATH = /app/runtimes 22:09:18:PREINSTALLED_MODULES_PATH = /app/preinstalled-modules 22:09:18:MODULES_PATH = /app/modules 22:09:18:PYTHON_PATH = /bin/linux/%PYTHON_RUNTIME%/venv/bin/python3 22:09:18:Data Dir = /etc/codeproject/ai 22:09:18:Server version: 2.2.4-Beta 22:09:18:Overriding address(es) 'http://+:32168, http://+:5000'. Binding to endpoints defined via IConfiguration and/or UseKestrel() instead. 22:09:21: 22:09:21:Module 'Object Detection (YOLOv5 6.2)' 1.6.1 (ID: ObjectDetectionYolo) 22:09:21:Module Path: /app/preinstalled-modules/ObjectDetectionYolo 22:09:21:AutoStart: True 22:09:21:Queue: objectdetection_queue 22:09:21:Platforms: all 22:09:21:GPU: Support enabled 22:09:21:Parallelism: 0 22:09:21:Accelerator: 22:09:21:Half Precis.: enable 22:09:21:Runtime: python38 22:09:21:Runtime Loc: Shared 22:09:21:FilePath: detect_adapter.py 22:09:21:Pre installed: True 22:09:21:Start pause: 1 sec 22:09:21:LogVerbosity: 22:09:21:Valid: True 22:09:21:Environment Variables 22:09:21:APPDIR = %CURRENT_MODULE_PATH% 22:09:21:CPAI_HALF_PRECISION = disable 22:09:21:CPAI_MODULE_SUPPORT_GPU = False 22:09:21:CUSTOM_MODELS_DIR = %CURRENT_MODULE_PATH%/custom-models 22:09:21:MODELS_DIR = %CURRENT_MODULE_PATH%/assets 22:09:21:MODEL_SIZE = Medium 22:09:21:USE_CUDA = False 22:09:21:YOLOv5_AUTOI