AscendKernelGen
/

KernelGen-LM-32B-RL

Model card Files Files and versions

AscendKernelGen commited on 15 days ago

Commit

9d5531e

·

verified ·

1 Parent(s): 5ce667d

Update README.md

Files changed (1) hide show

README.md +3 -0

README.md CHANGED Viewed

@@ -6,6 +6,9 @@ base_model:
 - Qwen/Qwen3-32B
 ---
 KernelGen-LM-32B is a state-of-the-art domain-adaptive large language model specialized for low-level NPU kernel generation, specifically for the Huawei Ascend architecture using the AscendC programming language. Built upon the Qwen3-32B backbone, it is trained on the Ascend-CoT dataset and refined via reinforcement learning with execution feedback. It achieves unprecedented success rates in generating complex, functional hardware kernels, improving compilation success on complex tasks from 0% (baseline) to 95.5% (Pass@10), while functional correctness achieves
 64.3% compared to the baseline’s complete failure.

 - Qwen/Qwen3-32B
 ---
+![License](https://img.shields.io/badge/License-Apache-yellow)
+[![arXiv](https://img.shields.io/badge/arXiv-2601.07160-b31b1b.svg)](https://arxiv.org/abs/2601.07160)
 KernelGen-LM-32B is a state-of-the-art domain-adaptive large language model specialized for low-level NPU kernel generation, specifically for the Huawei Ascend architecture using the AscendC programming language. Built upon the Qwen3-32B backbone, it is trained on the Ascend-CoT dataset and refined via reinforcement learning with execution feedback. It achieves unprecedented success rates in generating complex, functional hardware kernels, improving compilation success on complex tasks from 0% (baseline) to 95.5% (Pass@10), while functional correctness achieves
 64.3% compared to the baseline’s complete failure.