cuda-kernels

Write, benchmark, and integrate CUDA kernels for HuggingFace Diffusers and Transformers.

Updated Aug 27, 2026

Search Tags:#benchmark #gpu #cuda #kernels #huggingface #transformers #diffusers

Ready to run

npx skills add https://github.com/rdromer2/mambavision_tf_kernel --skill cuda-kernels-rdromer2

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: cuda-kernels
Source: https://github.com/rdromer2/mambavision_tf_kernel/tree/main/.agents/skills/cuda-kernels
Command: npx skills add https://github.com/rdromer2/mambavision_tf_kernel --skill cuda-kernels-rdromer2

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch, diffusers, transformers, kernels, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a structured approach to writing, benchmarking, and integrating optimized CUDA kernels for NVIDIA GPUs, specifically targeting HuggingFace Diffusers and Transformers workloads.

Core Features & Use Cases

Guidance for building high-performance CUDA kernels (RMSNorm, RoPE, GEGLU, AdaLN) and validating against PyTorch baselines.
Integration guidance with HuggingFace Kernels Hub (get_kernel) for loading pre-compiled kernels and diffusers/transformers pipelines.
Benchmarking templates and example workflows (video generation, micro-benchmarks) to compare optimized vs baseline performance.

Quick Start

Run the included scripts to benchmark and inject kernels into your Diffusers/Transformers workflow.

Frequently Asked Questions about cuda-kernels

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I write and benchmark custom CUDA kernels for HuggingFace Diffusers?▼

To write and benchmark custom CUDA kernels for HuggingFace Diffusers, use the structured workflow to develop high-performance kernels like RMSNorm or RoPE, validate them against PyTorch baselines, and compare optimized performance using benchmarking templates.

Can I integrate pre-compiled CUDA kernels from HuggingFace Hub into Transformers pipelines?▼

Yes, you can integrate pre-compiled CUDA kernels into Transformers pipelines using the HuggingFace Kernels Hub get_kernel guidance, which supports loading and injecting optimized kernels directly into your diffusers and transformers workflows.

What CUDA kernel operations does this benchmarking workflow support for Transformers?▼

This benchmarking workflow supports building and validating high-performance CUDA kernels for operations including RMSNorm, RoPE, GEGLU, and AdaLN, comparing their optimized performance against standard PyTorch baselines.

Do I need PyTorch and Diffusers installed to use these CUDA kernel integration scripts?▼

Yes, you need PyTorch, Diffusers, Transformers, and the kernels dependency installed to run the scripts, benchmark performance, and inject optimized CUDA kernels into your video generation and transformer workloads.

What is the best way to compare optimized CUDA kernels against standard PyTorch baselines?▼

The best way to compare optimized CUDA kernels against PyTorch baselines is using the provided benchmarking templates and example workflows, which enable micro-benchmarks and video generation performance comparisons.

Are there limitations when injecting custom CUDA kernels into Diffusers video generation pipelines?▼

Limitations involve end-to-end kernel development constraints and pipeline integration compatibility; validate your custom Diffusers CUDA kernels against PyTorch baselines first to ensure video generation outputs remain accurate before full deployment.