Download

Skill UI

Browse and discover 15726+ curated skills

All Development Artificial Intelligence Design & Creative Product & Business Data Science Marketing Soft Skills Productivity Engineering Languages

Search LLM Training , found 27 results

Default Newest Most Downloaded

GPU Curated Data Pipeline

Orchestra-Research/AI-Research-SKILLs

NeMo Curator streams GPU-accelerated curation for LLM training, with super-fast fuzzy deduplication, multimodal quality filters, semantic dedupl, PII redaction, NSFW checks, and RAPIDS-scaled pipelines for prepping massive corpora.

OpenRLHF Training Suite

openrlhf-training

Orchestra-Research/AI-Research-SKILLs

High-performance RLHF framework that combines Ray, vLLM, and ZeRO-3 to accelerate PPO/GRPO/RLOO/DPO training for 7B-70B+ models on distributed GPU clusters, providing a single-stack workflow for reward models and policy optimization.

LLM Quantization Toolkit

quantizing-models-bitsandbytes

Orchestra-Research/AI-Research-SKILLs

Quantizes HuggingFace LLMs to 8-bit or 4-bit with bitsandbytes, cutting memory by up to 75% while keeping accuracy, and supports QLoRA fine-tuning plus 8-bit optimizers for faster, memory-efficient training.

Production ML Engineering

senior-ml-engineer

alirezarezvani/claude-skills

Guides production ML engineering with patterns for deploying models, building MLOps pipelines, monitoring drift, integrating LLMs and RAG systems, and enforcing feature-store, retraining, and cost controls in live environments.

Simple Preference Optimization for LLM Alignment

Orchestra-Research/AI-Research-SKILLs

SimPO (Simple Preference Optimization) is a state-of-the-art, reference-free method designed for aligning Large Language Models (LLMs) using human preference data. It is an efficient alternative to DPO and PPO, notably outperforming DPO without requiring a separate reference model. It is ideal for practitioners who need faster, simpler, and more resource-efficient fine-tuning for preference alignment.

SLIME RL Training

slime-rl-training

Orchestra-Research/AI-Research-SKILLs

Guides post-training of GLM-class LLMs with Megatron-LM and SGLang, focusing on GRPO rollout generation, custom data buffers, and tight RL scaling workflows for research and production use.

Megatron Core LLM Training

training-llms-megatron

Orchestra-Research/AI-Research-SKILLs

Provides production-ready Megatron-Core workflows for training 2B–462B parameter LLMs with tensor/pipeline/expert/context parallelism, maximizing GPU efficiency on H100 clusters.

Volcano Engine RL Training

verl-rl-training

Orchestra-Research/AI-Research-SKILLs

Guides training large language models with Volcano Engine's verl library, covering PPO/GRPO workflows, distributed backends, multi-turn rollouts, reward configs, critic-enabled PPO, and production deployments using Ray workers.

Training LLMs on Hugging Face Cloud Infrastructure

hugging-face-model-trainer

sickn33/antigravity-awesome-skills

A comprehensive skill for fine-tuning and training large language and vision models using advanced techniques like TRL (SFT, DPO, GRPO) or Unsloth. This process runs entirely on managed Hugging Face cloud GPUs, eliminating the need for local GPU setup. It supports modern training workflows, model conversion (GGUF for local deployment), and ensures seamless artifact storage to the Hugging Face Hub.

PyTorch Training Recipes

ml-training-recipes

Orchestra-Research/AI-Research-SKILLs

Collection of battle-tested PyTorch training recipes covering architecture selection, optimizers, LR scheduling, precision tips, and debugging to train or fine-tune LLMs, vision, diffusion, and biomedical models with high GPU throughput.

CoreWeave Distributed GPU Training Workflow

coreweave-core-workflow-b

jeremylongshore/claude-code-plugins-plus-skills

This guide details how to set up and execute large-scale, distributed GPU training jobs on the CoreWeave platform. It covers both single-node multi-GPU setups and multi-node training utilizing PyTorch DDP and Kubernetes resources (A100/H100). This workflow is ideal for fine-tuning large language models (LLMs) or running complex deep learning models that require high-performance, shared-storage compute clusters.

CoreWeave GPU Performance Tuning Guide

coreweave-performance-tuning

jeremylongshore/claude-code-plugins-plus-skills

This guide provides expert strategies for optimizing GPU inference performance on CoreWeave infrastructure. It covers GPU selection based on workload (LLM, image generation, training), advanced techniques like continuous batching (using vLLM), autoscaling setup (HPA), and benchmarking data. Use this to maximize GPU utilization, minimize latency, and optimize large-scale AI model deployment.

Language