Understanding YOLO11's Deployment Options
ã¯ããã«
You've come a long way on your journey with YOLO11. You've diligently collected data, meticulously annotated it, and put in the hours to train and rigorously evaluate your custom YOLO11 model. Now, it's time to put your model to work for your specific application, use case, or project. But there's a critical decision that stands before you: how to export and deploy your model effectively.
This guide walks you through YOLO11's deployment options and the essential factors to consider to choose the right option for your project.
How to Select the Right Deployment Option for Your YOLO11 Model
When it's time to deploy your YOLO11 model, selecting a suitable export format is very important. As outlined in the Ultralytics YOLO11 Modes documentation, the model.export() function allows for converting your trained model into a variety of formats tailored to diverse environments and performance requirements.
çæ³çãªãã©ãŒãããã¯ãã¹ããŒããããŒããŠã§ã¢ã®å¶çŽãçµ±åã®ããããã®ãã©ã³ã¹ãåããªãããã¢ãã«ã®æå³ããéçšç¶æ³ã«ãã£ãŠç°ãªããŸãã次ã®ã»ã¯ã·ã§ã³ã§ã¯ãåãšã¯ã¹ããŒãã»ãªãã·ã§ã³ã詳ããèŠãŠããããããéžæããã¿ã€ãã³ã°ãç解ããŸãã
YOLO11's Deployment Options
Let's walk through the different YOLO11 deployment options. For a detailed walkthrough of the export process, visit the Ultralytics documentation page on exporting.
PyTorch
PyTorch is an open-source machine learning library widely used for applications in deep learning and artificial intelligence. It provides a high level of flexibility and speed, which has made it a favorite among researchers and developers.
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ïŒPyTorch ã¯ãã®äœ¿ãããããšæè»æ§ã§ç¥ãããŠãããããã®çµæãããå°éçã§æé©åãããä»ã®ãã¬ãŒã ã¯ãŒã¯ãšæ¯èŒããå Žåãçã®ããã©ãŒãã³ã¹ã«ãããŠè¥å¹²ã®ãã¬ãŒããªããçããå¯èœæ§ãããã
-
äºææ§ãšçµ±åïŒPython ã®æ§ã ãªããŒã¿ãµã€ãšã³ã¹ããã³æ©æ¢°åŠç¿ã©ã€ãã©ãªãšã®åªããäºææ§ãæäŸããŸãã
-
ã³ãã¥ããã£ã»ãµããŒããšãšã³ã·ã¹ãã ïŒæã掻æ°ã®ããã³ãã¥ããã£ã®1ã€ã§ãåŠç¿ããã©ãã«ã·ã¥ãŒãã£ã³ã°ã®ããã®è±å¯ãªãªãœãŒã¹ããããŸãã
-
ã±ãŒã¹ã¹ã¿ãã£ïŒç 究ãããã¿ã€ãã§ãã䜿ãããå€ãã®åŠè¡è«æã¯PyTorch ã§å±éãããã¢ãã«ãåç §ããŠããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒç©æ¥µçãªéçºãšæ°æ©èœã®ãµããŒãã«ããå®æçãªã¢ããããŒãã
-
ã»ãã¥ãªãã£ã«é¢ããèæ ®äºé ïŒã»ãã¥ãªãã£äžã®åé¡ã«å¯ŸããŠã¯å®æçã«ããããåœãŠãããã»ãã¥ãªãã£ã¯å°å ¥ç°å¢å šäœã«å€§ããå·Šå³ãããã
-
ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒã¢ãã«ã®ãã¬ãŒãã³ã°ãæšè«ãé«éåããããã«äžå¯æ¬ ãªãGPU ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³çšã®CUDA ããµããŒãã
TorchScript
TorchScript æ¡åŒµ PyTorchã®æ©èœã¯ãã¢ãã«ã®ãšã¯ã¹ããŒãã C++ ã©ã³ã¿ã€ã ç°å¢ã§å®è¡ã§ããããã«ããããšã§å®çŸã§ããŸããããã«ããã次ã®ãããªæ¬çªç°å¢ã«é©ããŠããŸãã Python ã¯äœ¿çšã§ããŸããã
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ïŒç¹ã«æ¬çªç°å¢ã§ã¯ããã€ãã£ãPyTorch ãããããã©ãŒãã³ã¹ãåäžãããããšãã§ããã
-
äºææ§ãšçµ±åïŒPyTorch ãã C++ ãããã¯ã·ã§ã³ç°å¢ãžã·ãŒã ã¬ã¹ã«ç§»è¡ã§ããããã«èšèšãããŠããŸãããäžéšã®é«åºŠãªæ©èœã¯å®å šã«ã¯ç§»è¡ã§ããªãå ŽåããããŸãã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒPyTorch ã®å€§èŠæš¡ã³ãã¥ããã£ã®æ©æµãåããããå°éçãªéçºè ã®ç¯å²ã¯çãã
-
ã±ãŒã¹ã¹ã¿ã㣠Python ã®ããã©ãŒãã³ã¹ã»ãªãŒããŒããããããã«ããã¯ãšãªã£ãŠããç£æ¥ç°å¢ã§åºã䜿çšãããŠããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒPyTorch ãäžè²«ããã¢ããããŒããå®æœã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒPython ãå®å šã«ã€ã³ã¹ããŒã«ãããŠããªãç°å¢ã§ãã¢ãã«ãå®è¡ã§ããããã«ããããšã§ãã»ãã¥ãªãã£ãåäžãããŸãã
-
ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒPyTorch ã®CUDA ãµããŒããç¶æ¿ããGPU ã®å¹ççãªå©çšãä¿èšŒããŸãã
ONNX
The Open Neural Network Exchange (ONNX) is a format that allows for model interoperability across different frameworks, which can be critical when deploying to various platforms.
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ïŒONNX ã¢ãã«ã¯ãå°å ¥ãããã©ã³ã¿ã€ã ã«ãã£ãŠããã©ãŒãã³ã¹ãå€ããå¯èœæ§ããããŸãã
-
äºææ§ãšçµ±åïŒãã¬ãŒã ã¯ãŒã¯ã«ãšããããªãæ§è³ªã«ãããè€æ°ã®ãã©ãããã©ãŒã ãããŒããŠã§ã¢ã«ãŸãããé«ãçžäºéçšæ§ã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒå€ãã®çµç¹ã«ãã£ãŠãµããŒããããå¹ åºããšã³ã·ã¹ãã ãšæé©åã®ããã®ããŸããŸãªããŒã«ã«ã€ãªãã£ãŠããã
-
ã±ãŒã¹ã¹ã¿ãã£ç°ãªãæ©æ¢°åŠç¿ãã¬ãŒã ã¯ãŒã¯éã§ã¢ãã«ã移åãããããã«é »ç¹ã«äœ¿çšããããã®æè»æ§ã瀺ããŠããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒãªãŒãã³ã¹ã¿ã³ããŒããšããŠãONNX ã¯å®æçã«ã¢ããããŒããããæ°ãããªãã¬ãŒã·ã§ã³ãã¢ãã«ããµããŒãããŠããã
-
ã»ãã¥ãªãã£ã®èæ ®ïŒä»ã®ã¯ãã¹ãã©ãããã©ãŒã ããŒã«ãšåæ§ã«ãå€æãšãããã€ã®ãã€ãã©ã€ã³ã§å®å šãªãã©ã¯ãã£ã¹ã確ä¿ããããšãäžå¯æ¬ ã§ãã
-
ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒONNX Runtimeã䜿çšãããšãã¢ãã«ã¯ããŸããŸãªããŒããŠã§ã¢æé©åã掻çšã§ããŸãã
OpenVINO
OpenVINO ã¯ãIntel ããŒããŠã§ã¢å šäœãžã®ãã£ãŒãã©ãŒãã³ã°ã¢ãã«ã®å±éã容æã«ããããã©ãŒãã³ã¹ãšã¹ããŒããåäžãããããã«èšèšãããIntel ããŒã«ãããã§ããã
-
ããã©ãŒãã³ã¹ãã³ãããŒã¯ïŒç¹ã«Intel CPUãGPUãVPUåãã«æé©åãããŠãããäºææ§ã®ããããŒããŠã§ã¢äžã§å€§å¹ ãªããã©ãŒãã³ã¹åäžãå®çŸããŸãã
-
äºææ§ãšçµ±åïŒIntel ã®ãšã³ã·ã¹ãã å ã§æãããæ©èœããããä»ã®ããŸããŸãªãã©ãããã©ãŒã ããµããŒãããŠããã
-
Community Support and Ecosystem: Backed by Intel, with a solid user base especially in the computer vision domain.
-
Case Studies: Often utilized in IoT and edge computing scenarios where Intel hardware is prevalent.
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒIntel ææ°ã®ãã£ãŒãã©ãŒãã³ã°ã¢ãã«ãšIntel ããŒããŠã§ã¢ããµããŒããããããOpenVINO ãå®æçã«ã¢ããããŒãããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒæ©å¯æ§ã®é«ãã¢ããªã±ãŒã·ã§ã³ã«é©ããå ç¢ãªã»ãã¥ãªãã£æ©èœãæäŸããŸãã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒå°çšã®åœä»€ã»ãããšããŒããŠã§ã¢æ©èœã掻çšããIntel ããŒããŠã§ã¢ã§ã®ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³çšã«èª¿æŽãããŠããŸãã
OpenVINO ã䜿çšããå±éã®è©³çŽ°ã«ã€ããŠã¯ãUltralytics Integration ã®ããã¥ã¡ã³ããåç §ããŠãã ããïŒIntel OpenVINO ãšã¯ã¹ããŒãã
TensorRT
TensorRT ã¯ãNVIDIA ãæäŸããé«æ§èœãªæ·±å±€åŠç¿æšè«ãªããã£ãã€ã¶ãšã©ã³ã¿ã€ã ã§ãã¹ããŒããšå¹çæ§ãå¿ èŠãšããã¢ããªã±ãŒã·ã§ã³ã«æé©ã§ãã
-
æ§èœãã³ãããŒã¯ïŒé«éæšè«ããµããŒãããNVIDIA GPUäžã§ãããã¯ã©ã¹ã®ããã©ãŒãã³ã¹ãå®çŸã
-
äºææ§ãšçµ±åïŒNVIDIA ããŒããŠã§ã¢ã«æé©ã§ããã®ç°å¢ä»¥å€ã§ã®ãµããŒãã¯éå®çã
-
ã³ãã¥ããã£ã»ãµããŒããšãšã³ã·ã¹ãã ïŒNVIDIA ã®éçºè ãã©ãŒã©ã ãããã¥ã¡ã³ããéãã匷åãªãµããŒããããã¯ãŒã¯ã
-
ã±ãŒã¹ã¹ã¿ãã£ãããªãç»åããŒã¿ã®ãªã¢ã«ã¿ã€ã æšè«ãå¿ èŠãšããæ¥çã§åºãæ¡çšãããŠããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒã:NVIDIA ã¯ãããã©ãŒãã³ã¹ãåäžãããæ°ããGPU ã¢ãŒããã¯ãã£ããµããŒãããããã«ãé »ç¹ãªã¢ããããŒãã§TensorRT ãç¶æããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒå€ãã®NVIDIA 補åãšåæ§ãã»ãã¥ãªãã£ã«éç¹ã眮ããŠããããå ·äœçãªå 容ã¯å°å ¥ç°å¢ã«äŸåããã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒNVIDIA GPUå°çšã«èšèšãããé«åºŠãªæé©åãšé«éåãæäŸããŸãã
CoreML
CoreML ã¯ã¢ããã«ã®æ©æ¢°åŠç¿ãã¬ãŒã ã¯ãŒã¯ã§ãiOS ãmacOSãwatchOSãtvOSãå«ãã¢ããã«ã®ãšã³ã·ã¹ãã ã«ããããªã³ããã€ã¹ã»ããã©ãŒãã³ã¹ã®ããã«æé©åãããŠããã
-
ããã©ãŒãã³ã¹ãã³ãããŒã¯ïŒApple補ããŒããŠã§ã¢ã®ããã€ã¹äžã§ã®ããã©ãŒãã³ã¹ã«æé©åãããããããªãŒã®æ¶è²»ãæå°éã«æããŸãã
-
äºææ§ãšçµ±åïŒAppleã®ãšã³ã·ã¹ãã å°çšã§ãiOS ãmacOSã¢ããªã±ãŒã·ã§ã³ã«åççãªã¯ãŒã¯ãããŒãæäŸããŸãã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒAppleã®åŒ·åãªãµããŒããšç±å¿ãªéçºè ã³ãã¥ããã£ãè±å¯ãªããã¥ã¡ã³ããšããŒã«ã
-
ã±ãŒã¹ã¹ã¿ãã£ã¢ããã«è£œåã®ããã€ã¹äžã§ã®æ©æ¢°åŠç¿æ©èœãå¿ èŠãšããã¢ããªã±ãŒã·ã§ã³ã§ãã䜿çšãããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒææ°ã®æ©æ¢°åŠç¿ã®é²æ©ãšAppleã®ããŒããŠã§ã¢ããµããŒãããããã«ãAppleã«ãã£ãŠå®æçã«æŽæ°ãããŸãã
-
Security Considerations: Benefits from Apple's focus on user privacy and data security.
-
ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒAppleã®ãã¥ãŒã©ã«ã»ãšã³ãžã³ãšGPU ãæ倧éã«æŽ»çšããæ©æ¢°åŠç¿ã¿ã¹ã¯ãå éã
TF SavedModel
TF SavedModel ã§ã TensorFlowã®æ©æ¢°åŠç¿ã¢ãã«ãä¿åããŠæäŸããããã®åœ¢åŒã§ãããç¹ã«ã¹ã±ãŒã©ãã«ãªãµãŒããŒç°å¢ã«é©ããŠããŸãã
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ãµãŒããŒç°å¢ã«ãããŠãç¹ã«TensorFlow Servingãšäœµçšããããšã§ãã¹ã±ãŒã©ãã«ãªããã©ãŒãã³ã¹ãæäŸããŸãã
-
äºææ§ãšçµ±åïŒã¯ã©ãŠãããšã³ã¿ãŒãã©ã€ãºã»ãµãŒããŒã®ãããã€ã¡ã³ããå«ããTensorFlow ã®ãšã³ã·ã¹ãã å šäœã«ãããå¹ åºãäºææ§ã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒTensorFlow ã®äººæ°ã«ãã倧èŠæš¡ãªã³ãã¥ããã£ã»ãµããŒãããããã€ãšæé©åã®ããã®èšå€§ãªããŒã«çŸ€ã
-
ã±ãŒã¹ã¹ã¿ãã£ãã£ãŒãã©ãŒãã³ã°ã¢ãã«ã倧èŠæš¡ã«æäŸããããã®æ¬çªç°å¢ã§åºã䜿çšãããŠããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒGoogle ãšTensorFlow ã³ãã¥ããã£ã®ãµããŒãã«ãããå®æçãªã¢ããããŒããšæ°æ©èœãä¿èšŒãããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒTensorFlow Servingã䜿çšãããããã€ã¡ã³ãã«ã¯ããšã³ã¿ãŒãã©ã€ãºã»ã°ã¬ãŒãã®ã¢ããªã±ãŒã·ã§ã³ã®ããã®å ç¢ãªã»ãã¥ãªãã£æ©èœãå«ãŸããŠããŸãã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒTensorFlow ã®ããã¯ãšã³ããéããŠæ§ã ãªããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ããµããŒãããŸãã
TF GraphDef
TF GraphDef ã¯ãã¢ãã«ãã°ã©ããšããŠè¡šçŸããTensorFlow ãããã¯ãéçãªèšç®ã°ã©ããå¿ èŠãªç°å¢ã«ãšã£ãŠæçãªãã©ãŒãããã§ããã
-
æ§èœãã³ãããŒã¯ïŒäžè²«æ§ãšä¿¡é Œæ§ã«éç¹ã眮ããéçèšç®ã°ã©ãã®å®å®ããããã©ãŒãã³ã¹ãæäŸããŸãã
-
äºææ§ãšçµ±åïŒTensorFlow ã®ã€ã³ãã©ã«ç°¡åã«çµ±åã§ããããSavedModel ãšæ¯ã¹ããšæè»æ§ã«æ¬ ããã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒéçã°ã©ããæé©åããããã®å€ãã®ãªãœãŒã¹ãå©çšå¯èœã§ãTensorFlow ã®ãšã³ã·ã¹ãã ã«ããåªãããµããŒãã
-
ã±ãŒã¹ã¹ã¿ãã£ïŒç¹å®ã®çµã¿èŸŒã¿ã·ã¹ãã ãªã©ãéçãªã°ã©ããå¿ èŠãªã·ããªãªã§åœ¹ç«ã€ã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒTensorFlow ã®ã³ã¢ã¢ããããŒããšäžŠè¡ããŠãå®æçãªã¢ããããŒããè¡ããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒTensorFlow ã®ç¢ºç«ãããã»ãã¥ãªãã£æ £è¡ã«ãããå®å šãªé åãä¿èšŒããã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒSavedModel ã»ã©æè»ã§ã¯ãªãããTensorFlow ã®ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ãªãã·ã§ã³ãå©çšã§ããã
TF ã©ã€ã
TF Lite㯠TensorFlowã®ã¢ãã€ã«ããã³çµã¿èŸŒã¿ããã€ã¹ã®æ©æ¢°åŠç¿åããœãªã¥ãŒã·ã§ã³ã§ãããããã€ã¹äžã®æšè«ã®ããã®è»œéã©ã€ãã©ãªãæäŸããŸãã
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ïŒã¢ãã€ã«æ©åšãçµã¿èŸŒã¿æ©åšã§ã®ã¹ããŒããšå¹çæ§ãè¿œæ±ããèšèšã
-
äºææ§ãšçµ±åæ§ïŒè»œéã§ãããããããŸããŸãªããã€ã¹ã§äœ¿çšå¯èœã
-
ã³ãã¥ããã£ã»ãµããŒããšãšã³ã·ã¹ãã ïŒGoogle ã匷åºãªã³ãã¥ããã£ããããéçºè åãã®ãªãœãŒã¹ãå¢ããŠããã
-
ã±ãŒã¹ã¹ã¿ãã£æå°éã®ãããããªã³ãã§ããã€ã¹äžã§ã®æšè«ãå¿ èŠãšããã¢ãã€ã«ã¢ããªã±ãŒã·ã§ã³ã§äººæ°ããããŸãã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãå®æçãªã¢ããããŒãã«ãããææ°ã®æ©èœãšã¢ãã€ã«ããã€ã¹åãã®æé©åãæäŸããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒãšã³ããŠãŒã¶ãŒããã€ã¹äžã§ã¢ãã«ãå®è¡ããããã®ã»ãã¥ã¢ãªç°å¢ãæäŸããŸãã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒGPU ãDSPãªã©ãããŸããŸãªããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ã»ãªãã·ã§ã³ããµããŒãã
TF ãšããžTPU
TF EdgeTPU ã¯ãGoogle ã® EdgeTPU ããŒããŠã§ã¢äžã§é«éãã€å¹ççãªã³ã³ãã¥ãŒãã£ã³ã°ãè¡ãããã«èšèšãããŠããããªã¢ã«ã¿ã€ã åŠçãå¿ èŠãšãã IoT ããã€ã¹ã«æé©ã§ãã
-
æ§èœãã³ãããŒã¯ïŒGoogle's EdgeTPU ããŒããŠã§ã¢ã§ã®é«éã§å¹ççãªã³ã³ãã¥ãŒãã£ã³ã°ã«æé©åãããŠããŸãã
-
äºææ§ãšçµ±åïŒEdgeTPU ããã€ã¹ã®TensorFlow Lite ã¢ãã«ã§ã®ã¿åäœããŸãã
-
ã³ãã¥ããã£ã»ãµããŒããšãšã³ã·ã¹ãã ïŒGoogle ããµãŒãããŒãã£ã®éçºè ãæäŸãããªãœãŒã¹ã§ãµããŒããæ¡å€§ã
-
ã±ãŒã¹ã¹ã¿ãã£äœã¬ã€ãã³ã·ãŒã§ãªã¢ã«ã¿ã€ã åŠçãå¿ èŠãªIoTæ©åšãã¢ããªã±ãŒã·ã§ã³ã«äœ¿çšã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒæ°ããEdgeTPU ããŒããŠã§ã¢ã»ãªãªãŒã¹ã®æ©èœã掻çšãããããç¶ç¶çã«æ¹è¯ã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒGoogle ãIoTããã³ãšããžããã€ã¹åãã®å ç¢ãªã»ãã¥ãªãã£ãšçµ±åã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒGoogle Coralããã€ã¹ãæ倧éã«æŽ»çšã§ããããã«ã«ã¹ã¿ã èšèšãããŠããŸãã
TF.js
TensorFlow.js(TF.js)ã¯ãæ©æ¢°åŠç¿æ©èœããã©ãŠã¶ã«çŽæ¥ããããã©ã€ãã©ãªã§ããããŠã§ãéçºè ãšãŠãŒã¶ãŒã«æ°ããªå¯èœæ§ãæäŸãããããã¯ãšã³ãã®ã€ã³ãã©ãå¿ èŠãšããã«ããŠã§ãã¢ããªã±ãŒã·ã§ã³ã«æ©æ¢°åŠç¿ã¢ãã«ãçµ±åããããšãã§ããŸãã
-
ããã©ãŒãã³ã¹ãã³ãããŒã¯ïŒã¯ã©ã€ã¢ã³ãããã€ã¹ã«ããããã劥åœãªããã©ãŒãã³ã¹ã§ããã©ãŠã¶ã§çŽæ¥æ©æ¢°åŠç¿ãå¯èœã
-
äºææ§ãšçµ±åïŒãŠã§ããã¯ãããžãŒãšã®é«ãäºææ§ã«ããããŠã§ãã¢ããªã±ãŒã·ã§ã³ã«ç°¡åã«çµ±åã§ããŸãã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒãŠã§ããšNode.jséçºè ã®ã³ãã¥ããã£ã«ãããµããŒããšãMLã¢ãã«ããã©ãŠã¶ã«ãããã€ããããã®æ§ã ãªããŒã«ã
-
ã±ãŒã¹ã¹ã¿ãã£ãµãŒããŒãµã€ãã®åŠçãå¿ èŠãšãããã¯ã©ã€ã¢ã³ããµã€ãã®æ©æ¢°åŠç¿ã®æ©æµãåããã€ã³ã¿ã©ã¯ãã£ããªãŠã§ãã¢ããªã±ãŒã·ã§ã³ã«æé©ã§ãã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãïŒãªãŒãã³ãœãŒã¹ã³ãã¥ããã£ããã®è²¢ç®ã«ãããTensorFlow ããŒã ãã¡ã³ããã³ã¹ãè¡ã£ãŠããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒãŠã§ãã»ãã©ãããã©ãŒã ã®ã»ãã¥ãªãã£ã»ã¢ãã«ãå©çšãããã©ãŠã¶ã®ã»ãã¥ã¢ã»ã³ã³ããã¹ãå ã§å®è¡ãããã
-
ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒWebGLã®ãããªããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ã«ã¢ã¯ã»ã¹ãããŠã§ãããŒã¹ã®APIã䜿çšããããšã§ãããã©ãŒãã³ã¹ãåäžãããããšãã§ããŸãã
PaddlePaddle
PaddlePaddle ã¯ãçŸåºŠãéçºãããªãŒãã³ãœãŒã¹ã®ãã£ãŒãã©ãŒãã³ã°ãã¬ãŒã ã¯ãŒã¯ã§ãããç 究è ã«ãšã£ãŠã¯å¹ççã§ãéçºè ã«ãšã£ãŠã¯äœ¿ããããããã«èšèšãããŠãããç¹ã«äžåœã§äººæ°ããããäžåœèªåŠçã«ç¹åãããµããŒããæäŸããŠããã
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯äœ¿ãããããšæ¡åŒµæ§ã«éç¹ã眮ãã競äºåã®ããããã©ãŒãã³ã¹ãæäŸã
-
äºææ§ãšçµ±åïŒBaiduã®ãšã³ã·ã¹ãã å ã«ããŸãçµ±åãããå¹ åºãã¢ããªã±ãŒã·ã§ã³ããµããŒãããŠããŸãã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒã³ãã¥ããã£ã¯äžççã«ã¯å°ããããç¹ã«äžåœã§ã¯æ¥éã«æé·ããŠããã
-
ã±ãŒã¹ã¹ã¿ãã£äžåœåžå Žããä»ã®äž»èŠãªãã¬ãŒã ã¯ãŒã¯ã®ä»£æ¿ãæ¢ããŠããéçºè ã«ãã䜿ãããŠããã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãäžåœèªã®AIã¢ããªã±ãŒã·ã§ã³ãšãµãŒãã¹ã®æäŸãäžå¿ã«å®æçã«æŽæ°ã
-
Security Considerations: Emphasizes data privacy and security, catering to Chinese data governance standards.
-
ããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒãã€ãã¥ç¬èªã®Kunlunããããå«ãæ§ã ãªããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ããµããŒãã
NCNN
NCNN ã¯ãã¢ãã€ã«ãã©ãããã©ãŒã åãã«æé©åãããé«æ§èœãªãã¥ãŒã©ã«ãããã¯ãŒã¯æšè«ãã¬ãŒã ã¯ãŒã¯ã§ããã軜éãã€å¹ççãªç¹ãç¹åŸŽã§ããªãœãŒã¹ãéãããŠããã¢ãã€ã«æ©åšãçµã¿èŸŒã¿æ©åšã«ç¹ã«é©ããŠããã
-
ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ïŒã¢ãã€ã«ã»ãã©ãããã©ãŒã åãã«é«åºŠã«æé©åãããŠãããARMããŒã¹ã®ããã€ã¹ã§å¹ççãªæšè«ãå¯èœã
-
äºææ§ãšçµ±åïŒæºåž¯é»è©±ãARMã¢ãŒããã¯ãã£ã®çµã¿èŸŒã¿ã·ã¹ãã ã§ã®ã¢ããªã±ãŒã·ã§ã³ã«é©ããŠããŸãã
-
ã³ãã¥ããã£ã®ãµããŒããšãšã³ã·ã¹ãã ïŒã¢ãã€ã«ããã³çµã¿èŸŒã¿MLã¢ããªã±ãŒã·ã§ã³ã«ç¹åããããããã ã掻çºãªã³ãã¥ããã£ã«æ¯ããããŠããã
-
ã±ãŒã¹ã¹ã¿ã㣠Android ããã®ä»ã®ARMããŒã¹ã®ã·ã¹ãã ã§ãå¹çãšã¹ããŒããéèŠãããã¢ãã€ã«ã»ã¢ããªã±ãŒã·ã§ã³ã«é©ããŠããŸãã
-
ã¡ã³ããã³ã¹ãšã¢ããããŒãæ§ã ãªARMããã€ã¹äžã§é«ãããã©ãŒãã³ã¹ãç¶æããããã«ç¶ç¶çã«æ¹è¯ãããŠããŸãã
-
ã»ãã¥ãªãã£ãžã®é æ ®ïŒããã€ã¹äžã§ããŒã«ã«ã«å®è¡ããããšã«éç¹ã眮ããããã€ã¹äžã§ã®åŠçã«åºæã®ã»ãã¥ãªãã£ã掻çšããã
-
ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ïŒARM CPUããã³GPUåãã«ã«ã¹ã¿ãã€ãºããããããã®ã¢ãŒããã¯ãã£ã«ç¹åããæé©åãæœãããŠããŸãã
Comparative Analysis of YOLO11 Deployment Options
The following table provides a snapshot of the various deployment options available for YOLO11 models, helping you to assess which may best fit your project needs based on several critical criteria. For an in-depth look at each deployment option's format, please see the Ultralytics documentation page on export formats.
å±éãªãã·ã§ã³ | ããã©ãŒãã³ã¹ã»ãã³ãããŒã¯ | äºææ§ãšçµ±å | å°å瀟äŒã®ãµããŒããšãšã³ã·ã¹ãã | ã±ãŒã¹ã¹ã¿ã㣠| ã¡ã³ããã³ã¹ãšã¢ããããŒã | ã»ãã¥ãªãã£ãžã®é æ ® | ããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ |
---|---|---|---|---|---|---|---|
PyTorch | åªããæè»æ§ã | Python ã©ã€ãã©ãªãšã®çžæ§ãæ矀 | è±å¯ãªãªãœãŒã¹ãšã³ãã¥ãã㣠| ç 究ãšãããã¿ã€ã | å®æçãã€ç©æ¥µçãªéçº | é åç°å¢ã«ãã | CUDA GPU ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³å¯Ÿå¿ |
TorchScript | ãããçç£ã«é©ããŠãããPyTorch | PyTorch ãã C++ ãžã®ã¹ã ãŒãºãªç§»è¡ | å°éæ§ã¯é«ããå¹ ã¯çãPyTorch | Python ãããã«ããã¯ãšãªã£ãŠããæ¥ç | äžè²«ããã¢ããããŒãPyTorch | å®å šã§ãªãã»ãã¥ãªãã£ã®åäžPython | ã®CUDA ãµããŒããç¶æ¿ããŠãããPyTorch |
ONNX | ã©ã³ã¿ã€ã ã«ããå€å | ç°ãªããã¬ãŒã ã¯ãŒã¯ã§é«ã | å€ãã®çµç¹ã«æ¯ããããå¹ åºããšã³ã·ã¹ãã | MLãã¬ãŒã ã¯ãŒã¯éã®æè»æ§ | æ°èŠäºæ¥ã®ããã®å®æçãªæŽæ° | å®å šãªå€æãšé åã®å®è·µ | æ§ã ãªããŒããŠã§ã¢ã®æé©å |
OpenVINO | Intel ããŒããŠã§ã¢ã«æé©å | Intel ãšã³ã·ã¹ãã å ã§æé« | ã³ã³ãã¥ãŒã¿ã»ããžã§ã³ã®åéã§ç¢ºåºããå®çžŸ | Intel ããŒããŠã§ã¢ã«ããIoTãšãšããž | Intel ããŒããŠã§ã¢ã®å®æçãªã¢ããããŒã | ç¹çŽ°ãªã¢ããªã±ãŒã·ã§ã³ã®ããã®å ç¢ãªæ©èœ | Intel ããŒããŠã§ã¢ã«å¯Ÿå¿ |
TensorRT | NVIDIA GPUã§ãããã¬ãã« | NVIDIA ããŒããŠã§ã¢ã«æé© | 匷åãªãããã¯ãŒã¯NVIDIA | ãªã¢ã«ã¿ã€ã æ åã»ç»åæšè« | æ°ããGPUã®ããã®é »ç¹ãªã¢ããããŒã | ã»ãã¥ãªãã£éèŠ | NVIDIA GPUçšã«èšèš |
CoreML | ããã€ã¹äžã®Apple補ããŒããŠã§ã¢ã«æé©å | ã¢ããã«ã®ãšã³ã·ã¹ãã å°çš | ã¢ããã«ãšããããããŒã®åŒ·åãªãµããŒã | ã¢ããã«è£œåã®ãªã³ããã€ã¹ML | ã¢ããã«ã®å®æçãªã¢ããããŒã | ãã©ã€ãã·ãŒãšã»ãã¥ãªãã£ã®éèŠ | ã¢ããã«ã®ãã¥ãŒã©ã«ã»ãšã³ãžã³ãšGPU |
TF SavedModel | ãµãŒããŒç°å¢ã§ã®æ¡åŒµæ§ | TensorFlow ãšã³ã·ã¹ãã ã«ãããå¹ åºãäºææ§ | TensorFlow 人æ°ã«ãã倧ããªæ¯æ | æš¡åãã¹ã±ãŒã«ã§æäŸ | Google ãšã³ãã¥ããã£ã«ããå®æçãªæŽæ° | äŒæ¥åãã®å ç¢ãªæ©èœ | æ§ã ãªããŒããŠã§ã¢ã»ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ |
TF GraphDef | éçèšç®ã°ã©ãã®å®å®æ§ | TensorFlow ã€ã³ãã©ãšããŸãçµ±å | éçã°ã©ããæé©åããããã®ãªãœãŒã¹ | éçã°ã©ããå¿ èŠãšããã·ããªãª | TensorFlow ã³ã¢ãšäžŠè¡ããŠã¢ããããŒã | 確ç«ãããTensorFlow ã»ãã¥ãªãã£æ £è¡ | TensorFlow å éãªãã·ã§ã³ |
TF ã©ã€ã | ã¢ãã€ã«ïŒçµã¿èŸŒã¿ã«ãããã¹ããŒããšå¹çæ§ | å¹ åºãããã€ã¹ã«å¯Ÿå¿ | 匷åºãªã³ãã¥ããã£ãGoogle | ãããããªã³ããæå°éã«æããã¢ãã€ã«ã¢ããªã±ãŒã·ã§ã³ | ã¢ãã€ã«åãææ°æ©èœ | ãšã³ããŠãŒã¶ãŒã»ããã€ã¹ã®ã»ãã¥ã¢ãªç°å¢ | GPU DSPãªã© |
TF ãšããžTPU | Google's EdgeTPU ããŒããŠã§ã¢ã«æé©åãããŠããŸãã | ãšããžå°çšTPU ããã€ã¹ | Google ããµãŒãããŒãã£ã®ãªãœãŒã¹ã§æé·ãã | ãªã¢ã«ã¿ã€ã åŠçãå¿ èŠãªIoTæ©åš | æ°ããEdgeTPU ããŒããŠã§ã¢ã®æ¹å | Googleå ç¢ãªIoTã»ãã¥ãªã㣠| Google ççã®ããã®ã«ã¹ã¿ã ãã¶ã€ã³ |
TF.js | é©åºŠãªãã©ãŠã¶å ããã©ãŒãã³ã¹ | ãŠã§ãæè¡ã«é«ãé¢å¿ | Webããã³Node.jséçºè ã®ãµããŒã | ã€ã³ã¿ã©ã¯ãã£ããªãŠã§ãã¢ããªã±ãŒã·ã§ã³ | TensorFlow ããŒã ãšã³ãã¥ããã£ãžã®è²¢ç® | ãŠã§ãã»ãã©ãããã©ãŒã ã®ã»ãã¥ãªãã£ã»ã¢ãã« | WebGLããã®ä»ã®APIã§åŒ·å |
PaddlePaddle | 競äºåã䜿ãããããæ¡åŒµæ§ | çŸåºŠã®ãšã³ã·ã¹ãã ãå¹ åºãã¢ããªã±ãŒã·ã§ã³ãµããŒã | äžåœãäžå¿ã«æ¥æé· | äžåœåžå ŽãšèšèªåŠç | äžåœã®AIã¢ããªã±ãŒã·ã§ã³ã«æ³šç® | ããŒã¿ã®ãã©ã€ãã·ãŒãšã»ãã¥ãªãã£ãéèŠ | ãã€ãã¥ã®åŽåŽããããå«ã |
NCNN | ã¢ãã€ã«ARMããŒã¹ã»ããã€ã¹ã«æé©å | ã¢ãã€ã«ããã³çµã¿èŸŒã¿ARMã·ã¹ãã | ãããã ã掻çºãªã¢ãã€ã«/çµã¿èŸŒã¿MLã³ãã¥ãã㣠| Android ããã³ARMã·ã¹ãã ã®å¹ç | ARMã®é«æ§èœã¡ã³ããã³ã¹ | ãªã³ããã€ã¹ã»ã»ãã¥ãªãã£ã®å©ç¹ | ARM CPUãšGPUã®æé©å |
ãã®æ¯èŒåæã§ã¯ããã€ã¬ãã«ãªæŠèŠã説æããŸããå°å ¥ã«ããã£ãŠã¯ããããžã§ã¯ãåºæã®èŠä»¶ãå¶çŽãèæ ®ããåãªãã·ã§ã³ã§å©çšå¯èœãªè©³çŽ°ãªããã¥ã¡ã³ãããªãœãŒã¹ãåç §ããããšãäžå¯æ¬ ã§ãã
ã³ãã¥ããã£ãšãµããŒã
When you're getting started with YOLO11, having a helpful community and support can make a significant impact. Here's how to connect with others who share your interests and get the assistance you need.
ããåºãã³ãã¥ããã£ãšã®é¢ãã
-
GitHub Discussions: The YOLO11 repository on GitHub has a "Discussions" section where you can ask questions, report issues, and suggest improvements.
-
Ultralytics DiscordãµãŒããŒ: Ultralytics ã«ã¯DiscordãµãŒããŒããããä»ã®ãŠãŒã¶ãŒãéçºè ãšäº€æµããããšãã§ããŸãã
å ¬åŒææžãšãªãœãŒã¹
- Ultralytics YOLO11 Docs: The official documentation provides a comprehensive overview of YOLO11, along with guides on installation, usage, and troubleshooting.
These resources will help you tackle challenges and stay updated on the latest trends and best practices in the YOLO11 community.
çµè«
In this guide, we've explored the different deployment options for YOLO11. We've also discussed the important factors to consider when making your choice. These options allow you to customize your model for various environments and performance requirements, making it suitable for real-world applications.
Don't forget that the YOLO11 and Ultralytics community is a valuable source of help. Connect with other developers and experts to learn unique tips and solutions you might not find in regular documentation. Keep seeking knowledge, exploring new ideas, and sharing your experiences.
ããããŒã»ãããã€ïŒ
ããããã質å
What are the deployment options available for YOLO11 on different hardware platforms?
Ultralytics YOLO11 supports various deployment formats, each designed for specific environments and hardware platforms. Key formats include:
- PyTorchç 究ããã³ãããã¿ã€ãã³ã°çšã§ãPython ã®çµ±åã«åªããŠããã
- TorchScriptPython ãå©çšã§ããªãæ¬çªç°å¢åãã
- ONNXã¯ãã¹ãã©ãããã©ãŒã ã®äºææ§ãšããŒããŠã§ã¢ã¢ã¯ã»ã©ã¬ãŒã·ã§ã³ã®ããã«ã
- OpenVINOIntel ãããŒããŠã§ã¢ã®ããã©ãŒãã³ã¹ãæé©åããã
- TensorRTNVIDIA GPUã§ã®é«éæšè«ã®ããã«ã
åãã©ãŒãããã«ã¯ç¬èªã®å©ç¹ããããŸãã詳ããã¯ããšã¯ã¹ããŒãã»ããã»ã¹ã®ããã¥ã¡ã³ããã芧ãã ããã
How do I improve the inference speed of my YOLO11 model on an Intel CPU?
To enhance inference speed on Intel CPUs, you can deploy your YOLO11 model using Intel's OpenVINO toolkit. OpenVINO offers significant performance boosts by optimizing models to leverage Intel hardware efficiently.
- Convert your YOLO11 model to the OpenVINO format using the
model.export()
é¢æ°ã§ããã - Intel OpenVINO Export ããã¥ã¡ã³ãã®è©³çŽ°ãªã»ããã¢ããã¬ã€ãã«åŸã£ãŠãã ããã
詳ããã¯ãããã°èšäºãã芧ãã ããã
Can I deploy YOLO11 models on mobile devices?
Yes, YOLO11 models can be deployed on mobile devices using TensorFlow Lite (TF Lite) for both Android and iOS platforms. TF Lite is designed for mobile and embedded devices, providing efficient on-device inference.
äŸ
ã¢ãã€ã«ãžã®ã¢ãã«å±éã®è©³çŽ°ã«ã€ããŠã¯ãTF Liteã€ã³ãã°ã¬ãŒã·ã§ã³ã¬ã€ãããåç §ãã ããã
What factors should I consider when choosing a deployment format for my YOLO11 model?
When choosing a deployment format for YOLO11, consider the following factors:
- ããã©ãŒãã³ã¹ïŒTensorRT ã®ãããªããã€ãã®ãã©ãŒãããã¯ãNVIDIA GPU äžã§åè¶ããé床ãæäŸããããOpenVINO ã¯Intel ããŒããŠã§ã¢çšã«æé©åãããŠããã
- äºææ§ïŒONNX ã¯ãããŸããŸãªãã©ãããã©ãŒã ã§å¹ åºãäºææ§ãæäŸããã
- çµ±åã®å®¹æãïŒCoreML ãTF Lite ã®ãããªãã©ãŒãããã¯ãããããiOS ãAndroid ã®ãããªç¹å®ã®ãšã³ã·ã¹ãã çšã«èª¿æŽãããŠããã
- Community Support: Formats like PyTorch and TensorFlow have extensive community resources and support.
æ¯èŒåæã«ã€ããŠã¯ã茞åºãã©ãŒãããã«é¢ããè³æããåç §ãã ããã
How can I deploy YOLO11 models in a web application?
To deploy YOLO11 models in a web application, you can use TensorFlow.js (TF.js), which allows for running machine learning models directly in the browser. This approach eliminates the need for backend infrastructure and provides real-time performance.
- Export the YOLO11 model to the TF.js format.
- ãšã¯ã¹ããŒãããã¢ãã«ãWebã¢ããªã±ãŒã·ã§ã³ã«çµ±åããŸãã
ã¹ããããã€ã¹ãããã®æé ã«ã€ããŠã¯ãTensorFlow.jsã®çµ±åã«é¢ããã¬ã€ããåç §ããŠãã ããã