Welcome to the Anakin GitHub.
Anakin is an cross-platform, high-performance inference engine, which is originally developed by Baidu engineers and is a large-scale application of industrial products.
Please refer to our release announcement to track the latest feature of Anakin.
Anakin supports a wide range of neural network architectures and diffrent hardware platform. It is easy to run Anakin at GPU/x86/ARM platform.
In order to giving full play to the performance of hardware, we optimize the forward prediction at diffrent levels.
Automatic graph fusion. The goal of all performance optimization under a given algorithm is to make ALU as busy as possible, Operator fusion can effectively reduce memory access and keep ALU busy.
Memory reuse. Forward prediction is a one-way calculation. We reuse the memory between the input and output of different operators, thus reducing the overall memory overhead.
Assembly level optimization. Saber is Anakin's underlying DNN library, which is deeply optimized at assembly level. Performance comparison between Anakin, TensorRT and Tensorflow-lite, please refer to the benchmark tests.
It is recommended to check out the Benchmark Readme
You might want to know more details of Anakin and make it better.
Python API is under-developing.
We appreciate your contributions!
You are welcome to submit questions and bug reports as Github Issues.
Copyright and License
Anakin is provided under the Apache-2.0 license.