Instructions to use unity/inference-engine-yolo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- unity-sentis
How to use unity/inference-engine-yolo with unity-sentis:
string modelName = "[Your model name here].sentis"; Model model = ModelLoader.Load(Application.streamingAssetsPath + "/" + modelName); IWorker engine = WorkerFactory.CreateWorker(BackendType.GPUCompute, model); // Please see provided C# file for more details
- Notebooks
- Google Colab
- Kaggle
issue with android phone.
#19
by 32nguyen - opened
For me, the model works fine on PC with a fixed texture for consistent testing.
On my Android phone, it also runs correctly with Vulkan + GPUCompute, but only at ~7 FPS. Using the CPU backend, it works as expected with ~17 FPS.
However, on OpenGLES, the results are incorrect with 28FPS — sometimes it detects objects with wrong labels, or the labels are correct but the bounding boxes glitch.
From my research, there seems to be a known issue with OpenGLES, but I’m not sure if it’s resolved in the latest version 2.5 that I’m using.
Could you let me know if you have successfully deployed it on android with GPU compute on OpenGLES? Thanks.