axinc-ai / ailia-models-cpp

C++ version of ailia models repository

ailia-models-cpp

The collection of pre-trained, state-of-the-art models for C++.

ailia models (Python version)

About ailia SDK

ailia SDK is a cross-platform high speed inference SDK. The ailia SDK provides a consistent C++ API on Windows, Mac, Linux, iOS, Android and Jetson. It supports Unity, Python and JNI for efficient AI implementation. The ailia SDK makes great use of the GPU via Vulkan and Metal to serve accelerated computing.

Install ailia SDK

Download ailia SDK

You can download a free evaluation version that allows you to evaluate the ailia SDK for 30 days. Please download from the trial link below.

https://ailia.jp/en/

Install ailia SDK

Copy the files located in the folder [ailia SDK]/library/ to the folder ./ailia/library/.

Install dependent libraries

Windows

gnumake and Visual Studio 2015 or newer are required. http://gnuwin32.sourceforge.net/packages/make.htm

Mac

Xcode Commandline Tools are required, they can be installed by running the command.

xcode-select --install

OpenCV is required, it can be installed by running the command.

brew install opencv.

Linux

OpenCV is required, it can be installed by running the command.

apt install libopencv-dev

Build

cd yolox
export AILIA_LIBRARY_PATH=../ailia/library
cmake .
make

Run

./yolox.sh -v 0

Supporting Models

We are now converting to C++. Please wait to complete conversion.

Audio processing

Name	Detail	Exported From	Supported Ailia Version
silero-vad	Silero VAD	Pytorch	1.2.15 and later
clap	CLAP	Pytorch	1.3.0 and later

Image classification

Name	Detail	Exported From	Supported Ailia Version
resnet50	Deep Residual Learning for Image Recognition	Chainer	1.2.0 and later
clip	CLIP	Pytorch	1.2.9 and later

Image segmentation

Name	Detail	Exported From	Supported Ailia Version
U-2-Net	U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection	Pytorch	1.2.2 and later

Object detection

Name	Detail	Exported From	Supported Ailia Version
yolov3-tiny	YOLO: Real-Time Object Detection	ONNX Runtime	1.2.1 and later
m2det	M2Det: A Single-Shot Object Detector based on Multi-Level Feature Pyramid Network	Pytorch	1.2.3 and later
yolox	YOLOX	Pytorch	1.2.6 and later

Pose estimation

Name	Detail	Exported From	Supported Ailia Version
lightweight-human-pose-estimation	Fast and accurate human pose estimation in PyTorch. Contains implementation of "Real-time 2D Multi-Person Pose Estimation on CPU: Lightweight OpenPose" paper.	Pytorch	1.2.1 and later

Face detection

Name	Detail	Exported From	Supported Ailia Version
yolov3-face	YOLO: Real-Time Object Detection	ONNX Runtime	1.2.1 and later

Face identification

Name	Detail	Exported From	Supported Ailia Version
arcface	pytorch implement of arcface	Pytorch	1.2.1 and later

Face recognition

Name	Detail	Exported From	Supported Ailia Version
face_alignment	2D and 3D Face alignment library build using pytorch	Pytorch	1.2.1 and later
mediapipe_iris	irislandmarks.pytorch	Pytorch	1.2.2 and later

Natural language processing

Name	Detail	Exported From	Supported Ailia Version
fugumt-en-ja	Fugu-Machine Translator	Pytorch	1.2.9 and later
fugumt-ja-en	Fugu-Machine Translator	Pytorch	1.2.10 and later
bert_maskedlm	huggingface/transformers	Pytorch	1.2.5 and later
sentence_transformers	sentence transformers	Pytorch	1.2.7 and later
t5_whisper_medical	error correction of medical terms using t5	Pytorch	1.2.13 and later

For NLP models

ailia.tokenizer is additionally required to use NLP models.

Download ailia.tokenizer

You can download a free evaluation version that allows you to evaluate the ailia.tokenizer. Please download from the trial link below.

https://axip-console.appspot.com/trial/terms/AILIA-TOKENIZER?lang=en

Install ailia SDK

Copy the files located in the folder [ailia Tokenizer]/library/ to the folder ./ailia_tokenizer/library/.

Build

cd fugumt-en-ja
export AILIA_LIBRARY_PATH=../ailia/library
export AILIA_TOKENIZER_PATH=../ailia_tokenizer/library
cmake .
make

Run

./fugumt-en-ja.sh

Build all samples

export AILIA_LIBRARY_PATH=../ailia/library
export AILIA_TOKENIZER_PATH=../ailia_tokenizer/library
cmake .
make