This document mainly introduces how to structure the NPU SDK.

You can clone the SDK from our Github and Gitlab.

$ mkdir workspace && cd workspace
$ git clone --recursive https://github.com/khadas/aml_npu_sdk.git

Enter the SDK directory,

$ cd {workspace}/aml_npu_sdk
$ ls
acuity-toolkit  android_sdk  Dockerfile  docs  LICENSE  linux_sdk  README.md

The SDK is mainly divided into several SDK, conversion tools and compilation tools, and docs.

acuity-toolkit    #Conversion tool , used to convert AI models
android_sdk       #Android SDK
docs              #Conversion related documents collection

Enter the Docs directory, DDK_Application_Guide_0.7.pdf describes each document. The document records a series of processes from conversion to integration, as well as some common problems.

acuity-toolkit is the conversion tool directory,

$ cd {workspace}/aml_npu_sdk/acuity-toolkit
$ ls
bin  demo  python  ReadMe.txt  requirements.txt

The main directory of interest is demo,

1. bin                   #Conversion is a collection of various tools used, most of which are not open source.
2. demo                  #Conversion script directory, convert AI model location
3. demo_hybird           #Mixed Input Conversion Tool
4. mulity_input_demo     #mulity input demo
5. python                #Used to convert the model and data corresponding to the python API
6. ReadMe.txt            #ReadMe.txt file explains how to convert and use
7. requirements.txt      #Conversion tool dependent environment

The environment dependency package required by the conversion tool can be installed directly on the PC or installed through the virtual environment virtualenv.

$ cd {workspace}/aml_npu_sdk/acuity-toolkit
$ cat requirements.txt
tensorflow==2.0.0
astor==0.8.0
numpy==1.18.0
scipy==1.1.0
Pillow==5.3.0
protobuf==3.11.2
networkx>=1.11
image==1.5.5
lmdb==0.93
onnx==1.6.0
h5py==2.10.0
flatbuffers==1.10
matplotlib==2.1.0
dill==0.2.8.2
ruamel.yaml==0.15.81
ply==3.11
torch==1.2.0

Among them, tensorflow==2.0.0 can be replaced by tensorflow==2.0.0a0.

The conversion script is in the `demo` directory,

$ cd {workspace}/aml_npu_sdk/acuity-toolkit/demo
$ ls
0_import_model.sh  1_quantize_model.sh  2_export_case_code.sh  data  extractoutput.py  inference.sh  model

Use scripts to convert AI models.

$ cd {workspace}/aml_npu_sdk/acuity-toolkit/demo
$ bash 0_import_model.sh && bash 1_quantize_model.sh && bash 2_export_case_code.sh 

After the conversion is completed, you can see the converted code in the xxxx_nbg_unify directory, here is the built-in model as an example.

$ cd {workspace}/aml_npu_sdk/acuity-toolkit/demo/mobilenet_tf_nbg_unify
$ ls
BUILD   makefile.linux   mobilenettf.vcxproj  vnn_global.h       vnn_mobilenettf.h   vnn_post_process.h  vnn_pre_process.h
main.c  mobilenet_tf.nb  nbg_meta.json        vnn_mobilenettf.c  vnn_post_process.c  vnn_pre_process.c

For the setting of conversion parameters, please refer to Model_Transcoding and Running User Guide Eng.pdf in Docs.