Neuralscope offers you benchmarking your AI solutions
For end-users:
We perform AI performance measurement on Android mobile devices by using NeuralScope Benchmark APP and we show computation capability comparisons on the  Ranking Info page . End-users can run the APP on their own mobiles and get the immediate cost-performance comparison on the mobile screen. Anyone is welcome to contribute performance data so that we can collect versatile comparisons from worldwide users.
For tech publishers:
Battery Life is critical for mobile devices. We provide an engineering edition APP for those technical reviewers or publishers to support in-house power measurements for your product evaluation. The " NeuralScope AI Benchmark Engineering Edition " is released based on a per-collaboration basis.  If you need related information, please Contact us.
For product providers:
NeuralScope APP is based on Android NNAPI to evaluate computational capability of mobile devices. There are four test categories, including object classification, object detection, object segmentation and multi-task hybrid mode (as below). Most test cases also contain two kinds of precision models, float model(FP 32) and quantized model(INT 8). Contact us  if your products have any problems to run the APP.
The deep learning based neural networks are able to recognize object classes for one or more given input photos. They can recognize 1000 different object classes. There are four models, mobilenet-V1, mobilenet-V2, Resnet-50, and Inception-V3, in our benchmarking App. Individually, we provide one float model and one quantized model for each network.
Now, you can perform object counting on your phone. The model we use is a combination of mobilenet, a light-weight classification model, and single shot multibox detector(SSD), an object detector does not resample pixels or feature maps for bounding box hypotheses, can detect 80 different object classes. It improves in speed for high-accuracy detection.
Image Segmentation
The task evaluates whether you can perform auto background removing for exchange scene on your phone. It can recognize 20 different object classes and segment the recognized object using different colors. The model is resnet-50 with atrous convolution layers embedded.