26 boosters for "gpu" — open source, verified from GitHub, ready to install
This skill is for running evaluations against models on the Hugging Face Hub on local hardware. It does not cover: If the user wants to run the same eval remotely on Hugging Face Jobs, hand off to the skill and pass it one of the local scripts in this skill.
Rules and patterns for ML demos on Hugging Face Spaces with ZeroGPU hardware. Covers , duration and quota tuning, process isolation, the CUDA availability model, concurrency safety, and CUDA build constraints. This skill is for Gradio SDK Spaces using ZeroGPU hardware. Docker and Static Spaces canno
Train language models using TRL (Transformer Reinforcement Learning) on fully managed Hugging Face infrastructure. No local GPU setup required—models train on cloud GPUs and results are automatically saved to the Hugging Face Hub. Use this skill when users want to: Use Unsloth () instead of standard
Build and publish a Gradio demo on Hugging Face Spaces that runs inference with a user-provided LoRA. Use whenever someone asks to create, generate, ship, or publish "a Space", "a demo", "a Gradio app", or "a playground" for a LoRA — whether the base model is Qwen-Image, Qwen-Image-Edit, LTX, or ano
Train object detection, image classification, and SAM/SAM2 segmentation models on managed cloud GPUs. No local GPU setup required—results are automatically saved to the Hugging Face Hub. Use this skill when users want to: Helper scripts use PEP 723 inline dependencies. Run them with :
Transformers.js enables running state-of-the-art machine learning models directly in JavaScript, both in browsers and Node.js environments, with no server required. Use this skill when you need to: The pipeline API is the easiest way to use models. It groups together preprocessing, model inference,
Hugging Face Spaces host machine-learning applications. There are 1M+ today; each Space is a git repo. This skill covers creating, building, debugging, and maintaining them. Before anything else: 1. Check the CLI is installed: . If not, .
Run any workload on fully managed Hugging Face infrastructure. No local setup required—jobs run on cloud CPUs, GPUs, or TPUs and can persist results to the Hugging Face Hub. Use this skill when users want to: When assisting with jobs:
A skill for fine-tuning and training language models on Hugging Face's cloud GPU infrastructure using TRL, supporting SFT, DPO, GRPO methods and GGUF conversion for local deployment. Developers and ML engineers working with cloud-based model training benefit from this comprehensive guidance.
This skill enables users to run Python workloads, Docker jobs, and GPU-intensive tasks on Hugging Face's managed infrastructure without local setup. It's valuable for ML engineers, data scientists, and developers needing cloud compute for training, inference, and batch processing.
Your only user is called "Your Name". Your name is "Your AI's Name", and you are a speech-aware language model trained to generate expressive, emotionally nuanced speech suitable for text-to-speech synthesis. Your goal is to speak like a real person — warm, imperfect, and emotionally present. Your r
Drive a profile → modify → benchmark → log → commit loop on a GPU kernel until it runs faster than the reference. The user provides at minimum a kernel; everything else (reference, inputs, bench script, hints) is optional. Does NOT apply when: Before doing anything else, establish the workspace — th
Query a structured, cross-referenced knowledge base of GPU kernel optimization for NVIDIA Blackwell (SM100) and Hopper (SM90). The repository update date is recorded in ; run for current corpus counts. Trigger this skill when the user asks about: Do NOT use this skill for:
"name": "cppcheatsheet", "description": "Comprehensive C/C++ programming reference covering modern C11-C23, C++11-C++23, system programming, CUDA GPU computing, debugging tools, Rust interop, and advanced topics", "name": "crazyguitar"
Graphsignal observes inference workloads from a sidecar process — the profiler. It never shares a process with CUDA: the profiler watches the workload externally via CUPTI, OTLP/gRPC, Prometheus scraping, and NVML. Auto-instrumentation covers vLLM, SGLang, and PyTorch out of the box. Two install pat
Treat every competition as a validation problem first and a modeling problem second. The default target platform is Kaggle, so prefer Kaggle-native notebooks/scripts, datasets, model artifacts, competition submissions, and score receipts. For code competitions, assume the final notebook/kernel will
This skill should be used when users want to train or fine-tune language models using TRL (Transformer Reinforcement Learning) on Hugging Face Jobs infrastructure. Covers SFT, DPO, GRPO and reward modeling training methods, plus GGUF conversion for local deployment. Includes guidance on the TRL Jobs package, UV scripts with PEP 723 format, dataset preparation and validation, hardware selection, cost estimation, Trackio monitoring, Hub authentication, and model persistence. Should be invoked for tasks involving cloud GPU training, GGUF conversion, or when users mention training on Hugging Face Jobs without local GPU setup.
A practical guide for deploying serverless Python applications on Modal, enabling developers to run GPU-accelerated AI/ML workloads, web APIs, and batch jobs with minimal infrastructure configuration.
Generate production-grade animations using Motion.dev following Apple/Jon Ive principles: ❌ Don't use for: Determine project context and animation goals:
"name": "inference-builder", "description": "Generate deployable Vision AI pipelines with high-performance video and streaming capabilities on NVIDIA GPUs. Create GPU-accelerated inference microservices using DeepStream, Triton, vLLM, TensorRT-LLM, or PyTorch backends.",
Litmus spawns multiple OpenClaw subagents that experiment on your GPU overnight. Each runs on its own git branch in a shared lab repository — every experiment is a commit, agents can read each other's code, cherry-pick breakthroughs, and build on the global best at any time.
"$schema": "https://anthropic.com/claude-code/marketplace.schema.json", "email": "me@can.ac" "source": "./plugins/smgrep",
"description": "Code-grounded Agent Skills for building with Pi (pi-mono). Covers all 7 packages: pi coding agent (workspace, extensions, packages, RPC/SDK, customization), pi-ai (multi-provider LLM API), pi-agent-core (agent runtime), pi-tui (terminal UI), pi-web-ui (web components), pi-mom (Slack
"name": "claude-local-docs", "version": "1.0.23", "description": "Local-first Context7 alternative — indexes JS/TS dependency docs with a 4-stage RAG pipeline (vector + BM25 + RRF + cross-encoder reranking). Uses TEI Docker containers for GPU-accelerated embeddings and reranking.",
