AscendKernelGen/KernelGen-LM-32B-RL
TEXT GENERATIONConcurrency Cost:2Model Size:32BQuant:FP8Ctx Length:32kPublished:Jan 23, 2026Architecture:Transformer Cold

KernelGen-LM-32B-RL by AscendKernelGen is a 32 billion parameter domain-adaptive large language model built on the Qwen3-32B backbone, specialized for low-level NPU kernel generation for the Huawei Ascend architecture using AscendC. It is trained on the Ascend-CoT dataset and refined with reinforcement learning using execution feedback. This model achieves high success rates in generating complex, functional hardware kernels, significantly improving compilation success and functional correctness for NPU programming tasks.

Loading preview...