An interactive Ascend-NPU process viewer
-
Updated
May 6, 2026 - Python
An interactive Ascend-NPU process viewer
A general-purpose real-time streaming media and deep learning inference acceleration framework, supporting H264, H265, AAC, MP4, FLV, RTSP, RTMP, and YOLO.实时流媒体及深度学习推理加速通用处理框架,支持H264、H265、AAC、MP4、FLV、RTSP、RTMP、YOLO。
RAG for visually rich documents — layout, formula and table structure survive the pipeline. Ported from CUDA to Huawei Ascend NPUs and tested on Ascend 910C.
AscendMate · 昇腾部署易用一指禅 —— 面向昇腾智算服务器的环境搭建、模型微调、推理部署、算子开发手册
Ascend Assistant · 昇腾服务器助手 —— Agent Skill,帮助用 AI 操作/查询/排障昇腾智算服务器(环境检测/排障/命令生成/部署引导/性能调优),与 AscendMate 手册深度联动
Ascend NPU fork of nanochat for LLM training with torch_npu/HCCL (experimental)
Native AscendC Mamba2 selective scan / SSD forward-backward custom operator for Huawei Ascend 910B3 and 950PR, with CANN, torch_npu, A100 benchmarks and msprof profiling.
Out-of-tree HCCL Task DAG and P2P traffic capture studio
Automated GLM interaction detection via CANN, NID scores, and SHAP interaction values for insurance pricing.
Qwen3.8-27B on original Ascend 910: custom CANN kernels, six-NPU experiments, measured decode speedups and reproducible build recipes.
Cholesky decomposition reference implementation on Ascend NPU
Predicting California house prices with MindSpore
A repository to store neural network test code.
Linear classification of CIFAR-10 dataset with MindSpore
To associate your repository with the cann topic, visit your repo's landing page and select "manage topics."