Ascend competition: CAN in Mindspore

Ramalama is an open-source tool that simplifies the local use and serving of AI models for inference from any source through the familiar approach of containers.

MindX DL 组件部署/ 静态虚拟化、动态虚拟化 6.0.RC2

MindXDL Operator project is to manage the infrastructure required for Ascend in a containerized manner. This involves components such as firmware, drivers, runtime, device-plugin, nodeD, export, and ascend-volcano

该项目由于部署在Huawei Ascend 310上由于推理运算，使用yolov5s模型，结合roslibpy与rosbridge进行消息交互，实现ROS系统。

Path Tracing based on huawei ascend npu and written by c++ and ascend-c

A common package for Huawei pyACL application development.

Add Ascend support for OpenRLHF

Using models to avoid the old fall down

conv accelerate 基本的卷积加速算法

Deeplabv3+ om model inference program on the Huawei Ascend platform

This example demonstrates a low-level native sort with duplicate elimination. It builds array (a) from values in (b), then reconstructs (b) with unique sorted values. Runtime is n+m, where m is the maximum element. Implemented in Python, MATLAB, and JavaScript.

基于自编码器的健康度实时预测模型，并在昇腾Atlas 300I Duo上实现国产化部署

昇腾加速卡驱动（rust binding）

Implement the GroupNorm operator using Ascend C.

benchmark gpu/npu flops and bandwidth

Implement the SovleEx operator using Ascend C.

A step-by-step guide to fine-tuning the Qwen3-8B model on Huawei Ascend via AutoDL. 一份在AutoDL平台使用华为昇腾（Ascend）对Qwen3-8B模型进行微调和推理的保姆级教程。

Porting Caffe, a deep learning framework, to HUAWEI Atlas 200 DK.

implement a LLM