NN Model Conversion Steps using Acuity Toolkit on Docker

Step 1: Pre-import Preparation

  • Make sure that your have installed and setup Docker before continuing to the following steps, kindly refer to the installation page if you have not done so.

  • Please verify that the following command is added in ‘pegasus_export_ovx.sh’ to ensure successful export on Docker.

--optimize 'VIP8000NANONI_PID0XAD' \
--pack-nbg-unify \
--viv-sdk '/opt/acuity/Vivante_IDE/VivanteIDE5.8.1.1/cmdtools' \

  • To import a model into Acuity, please place the model in a folder which is named after it and make sure that it is accessible by the import script.

  • If you do not have a customized model ready, you may download the sample models provided on our Repository for testing. Do note that these models are for testing and development purposes only.

  • In the folder containing model, create a file named dataset.txt, list down the accessible paths to your images there, you may choose to place all images within the same folder.

Step 2: Import Model

Import your model using command ./pegasus_import.sh <model_name>

Once import is successful, kindly refer to this page to modify the inputmeta.yml according to the type of imported model.

Step 3: Quantize Model

Perform quantization process using command ./pegasus_quantize.sh <model_name> <quantization_type>, where quantization_type can be uint8 or int16.

Step 4: Export Model

Once quantization process is completed, please export your quantized model using command ./pegasus_export_ovx.sh <model_name> <quantization_type>.

If the export is successful, you will be able to find your ‘.nb’ converted model contained in the ‘nbg-unify’ folder.