atlascollege / models-1

Provides the DaVinci models that are converted from a mainstream, open source, deep learning network models

EN|CN

Introduction

The models repository provides the network model used in the ascenddk sample. The network model is the Davici model converted from the mainstream open-source neural learning network.

Model Conversion Guide

Convert the open source network model into the Davici model supported by Huawei NPU chips by referring to the following guide.

Offline Model Conversion Guide

Model List

Model Name	Description	Model Path
face_detection.om	Network model for face detection	computer_vision/object_detect/face_detection
vanillacnn.om	Network model for marking facial feature points	computer_vision/classification/vanillacnn
sphereface.om	Network model for obtaining feature vectors	computer_vision/classification/sphereface
car_color.om	Network model for identifying the vehicle color	computer_vision/classification/car_color
car_type.om	Network model for identifying the vehicle brand	computer_vision/classification/car_type

Note

Submit the model based on the README_MODEL_TEMPLATE.md requirements. The model in this warehouse is an open source network model obtained from the Internet, which is not optimized by Huawei. The accuracy and performance depend on the source network. If do not meet your requirements, contact Huawei to obtain a commercial solution.