-
(KernelBook)[https://huggingface.co/datasets/GPUMODE/KernelBook] is a curated collection of pairs of pytorch programs and equivalent triton code (generated by torch inductor)
-
(KernelBench)[https://huggingface.co/datasets/ScalingIntelligence/KernelBench] A benchmark designed to evaluate the ability of LLMs to generate efficient GPU kernels for optimizing neural network performance
-
(ComputeEval)[https://github.com/nvidia/compute-eval] ComputeEval: Evaluating Large Language Models for CUDA Code Generation
-
(Kevin-32B)[https://cognition.ai/blog/kevin-32b] Kevin-32B: Multi-Turn RL for Writing CUDA Kernels
-
(KernelLLM)[https://huggingface.co/facebook/KernelLLM] a large language model based on Llama 3.1 Instruct, which has been trained specifically for the task of authoring GPU kernels using Triton.
-
(Kwen0.1)[https://huggingface.co/cdreetz/kwen2.5-1.5b] Qwen finetuned to write triton kernels
-
(KwenSFTDataset)[https://huggingface.co/datasets/cdreetz/triton-sft-dataset] corresponding dataset used to train Kwen