A comprehensive knowledge base for Huawei Ascend NPU development, structured as distributed Agent Skills. https://ascend-ai-coding.github.io/awesome-ascend-skills/
A comprehensive knowledge base for Huawei Ascend NPU development, structured as distributed AI Agent Skills. Includes npu-smi device management, HCCL performance testing, and ATC model conversion.
A starter package for Huawei Ascend development that includes device checks, container support, and basic PyTorch NPU capabilities. It is intended as the first setup step for general projects.
An installation package for Ascend model conversion, quantization, inference deployment, and evaluation. It includes tools and guides for vLLM-Ascend, ais-bench, Diffusers, and related model workflows.
A toolkit for AI for Science work on Huawei Ascend processors, covering profiling, model migration, and TensorFlow or Keras workflow choices. AI for Science means using machine learning to study scientific problems.
A set of skills for running Diffusers, a library for image and other generative AI models, on Huawei Ascend NPU processors. It covers environment setup, model weights, and pipeline inference.
A research tool that collects developer feedback from the Huawei Ascend community forum and identifies usability problems, failures, and error reports within a chosen time range.
AI for Science skill set for Huawei Ascend NPU. Includes a main routing skill for profiling collection, model migration, and TensorFlow route selection, plus specialized leaf skills for Boltz2, BoltzGen, DeepFRI, DiffSBDD, GENERator, OligoFormer, ProteinBERT, TensorFlow Community deployment, TF-to-PyTorch conversion…
A pair of code-review skills for Ascend C, Huawei’s language and development environment for Ascend NPU operators, with code-review rules and API documentation lookup.
A collection of 91 skills for working with Huawei Ascend AI processors, including code analysis, research-paper discovery, inference repositories, and model migration.
A collection of nine skills for analysing Huawei Ascend performance data and improving related development workflows.
★not rated 173 yesterdayA
tokens not measured
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: