Age & Gender Classifier Pricing
The age and gender classifier is distributed as a binary model in TensorFlow format (single protobuf TensorFlow model format and SavedModel format for TensorFlow Serving). The TensorFlow models can be converted and deployed on different inference engines, like OpenVINO, NVIDIA TensorRT, TensorFlow Lite, etc. There are examples for using the age and gender prediction with source code in Python and C++, which run under Windows and Linux.
We provide free REST API for online testing of the accuracy. If you would like to evaluate our classifiers, we will give you access to the server and the API protocol.
The classifier has several versions with different input sizes: 224px; 128px; 96px; 64px. The smaller the input size, the faster it runs, at the expense of lower accuracy. The version with the input size of 224x224 is the most accurate, and the 64x64 is the fastest one.
The pricing scheme is based on the number of licenses. Each license is for one device (CPU, vCPU or GPU). The licenses allow integration of the age and gender recognition model into software and hardware products. When the classifier is used on the cloud, the number of licenses have to be equal to the number of vCPU's or GPU's that run the model. When used on edge devices, the number of licenses have to be equal to the number of the devices. The licenses are perpetual (not limited in time). The updates are optional and are released several times a year. When new versions of the models are available, they can be purchased at 20% of the license price.
||Gender prediction accuracy (%)
||Age estimation mean absolute error (in years)
224x224 Model Pricing
|Number of Licenses *
||Price per License
||Price of new versions
||Time and usage limitations
|10 - 49
|50 - 99
|100 - 499
|500 - 999
* The minimum order quantity is 10 licenses (€3000).