Performance-optimized AI inference on your GPUs. Unlock superior throughput by selecting and tuning engines like vLLM or SGLang.

an edge-real-time anchor-free object detector with decent performance

Ascend PyTorch adapter (torch_npu). Mirror of https://gitcode.com/Ascend/pytorch

Persist and reuse KV Cache to speedup your LLM.

An interactive Ascend-NPU process viewer

Deep Gradient Network for Camouflaged Object Detection (MIR 2023). Our codebase supports PyTorch & Jittor & Huawei-Ascend platforms..

ChineseHarm-Bench: A Chinese Harmful Content Detection Benchmark

YOLOv5 om model inference program on the Huawei Ascend platform

在Ascend服务器上使用Dcoker创建指定的可以运行的容器

AI/DL enabled UAV platform powered by the Atlas 200 DK and Ascend Ecosystem

NumPy & SciPy for GPU as well as Ascend NPU (80% math API supported, alpha)

benchmark gpu/npu flops and bandwidth

Implement the SovleEx operator using Ascend C.

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

Using models to avoid the old fall down

Deeplabv3+ om model inference program on the Huawei Ascend platform

Implement the GroupNorm operator using Ascend C.

A DevOps toolkit for Huawei Ascend NPU environments that automates Docker environment configuration and resolves CANN/driver/framework version compatibility issues.

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.