On June 30, Huawei officially announced the open-sourcing of the PanGu 7B parameter dense model, the PanGuPro MoE 72B parameter mixture-of-experts model, and the model inference technology based on Ascend.
This open-source initiative is a key step in Huawei's strategy to build an Ascend ecosystem, aiming to promote research and innovation in large model technologies, accelerate the application of artificial intelligence across industries, and create value.
Currently, the model weights of PanGuPro MoE 72B, the basic inference code, and the inference code for ultra-large-scale MoE models based on Ascend have been officially launched on the open-source platform. The model weights and inference code for PanGu 7B will also be available soon.
Huawei sincerely invites global developers, enterprise partners, and researchers to download and use the relevant resources, and provide feedback to jointly improve the related technologies. Those who need more information can visit https://gitcode.com/ascend-tribe.