Topic · Automation and Agents
Fine-Tuning and Model Training skills for Claude Code and Codex
Skills for fine-tuning and training models: SFT, DPO and GRPO with TRL, LoRA and QLoRA with PEFT, Axolotl, Unsloth and LLaMA-Factory, distributed training with DeepSpeed, FSDP and Megatron, NVIDIA NeMo and TAO recipes for language and vision models, and managed tuning on Google's Agent Platform and Microsoft Foundry. They encode the configuration details that decide whether a run converges.
Highest-scoring model training skills in the October 22, 2026 snapshot: tao-validate-recipe-transfer from NVIDIA/skills (88/100), train-sentence-transformers from huggingface/skills (83/100) and knowledge-distillation from Orchestra-Research/AI-Research-SKILLs (77/100). Scores come from the Suede Skill Check.
- 127 listed skills
- 7 repositories
- 9 with full pages
88 from official publishers. 23 bundle scripts. 4 connect to an outside service. Listed in AI Agents and LLM Engineering (120), Science and Research (6), DevOps and Cloud (1).
All 127 model training skills
No matching skills. Try fewer words or clear your search.
| Skill | Repository | What it does |
|---|---|---|
| huggingface-lora-space-builder 75/100 · C | huggingface/skills | Builds and publishes a Gradio demo on Hugging Face Spaces for a user-provided LoRA, including the base pipeline and a task-specific UI. |
| knowledge-distillation 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Compresses large language models into smaller student models using knowledge distillation, covering temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training. |
| ml-training-recipes 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Provides PyTorch training recipes for LLMs, vision, diffusion, and genomics, covering training loops, optimizers, learning rate schedules, mixed precision, debugging, and GPU throughput. |
| model-merging 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Merges multiple fine-tuned models with mergekit to combine capabilities without retraining, using methods such as SLERP, TIES-Merging, DARE, and Task Arithmetic. |
| model-pruning 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Reduces LLM size and speeds inference with pruning techniques like Wanda and SparseGPT, without retraining, including structured, unstructured, N:M, and magnitude pruning. |
| sentence-transformers 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Generates sentence, text, and image embeddings with 5000+ pre-trained models for semantic similarity, clustering, retrieval, RAG, and semantic search. |
| sentencepiece 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Guides training language-independent SentencePiece tokenizers with BPE or Unigram directly on raw text, for multilingual, CJK, and reproducible tokenization needs. |
| tao-validate-recipe-transfer 88/100 · B | NVIDIA/skills | Ports a published computer vision paper's code and training recipe onto a customer's dataset, and diagnoses why such transfers give bad results. |
| train-sentence-transformers 83/100 · B | huggingface/skills | Trains and fine-tunes sentence-transformers models, including bi-encoders, cross-encoders, sparse encoders, and multi-vector encoders, with guidance on loss selection, hard-negative mining, evaluators, and Hub publishing. |
| agent-platform-tuning 75/100 · C | google/skills | Agent Platform Model Tuning. Use when you need to fine-tune open models or Gemini models using Agent Platform infrastructure. Don't use for model… |
| agent-platform-tuning-management 84/100 · B | google/skills | Manages GenAI tuning jobs in Agent Platform. Use this to list, get, or cancel ongoing model tuning jobs. Don't use for fine-tuning models (use… |
| axolotl 67/100 · D | Orchestra-Research/AI-Research-SKILLs | Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support |
| constitutional-ai 74/100 · C | Orchestra-Research/AI-Research-SKILLs | Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with self-critique/revision, then… |
| deepchem 81/100 · B | K-Dense-AI/claude-scientific-skills | Builds molecular property prediction and MoleculeNet workflows with DeepChem, including SMILES featurization, scaffold or grouped holdouts, masked… |
| deepspeed 61/100 · D | Orchestra-Research/AI-Research-SKILLs | Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention |
| digital-health-clinical-asr-finetune 87/100 · B | NVIDIA/skills | Stage 4 of the Clinical ASR Flywheel. Use when priority KER is above 0.3 to run stock NeMo SFT on Parakeet TDT v2 and offline cycle N+1 re-eval. NOT… |
| distributed-llm-pretraining-torchtitan 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1… |
| fine-tuning-expert 84/100 · B | jeffallan/claude-skills | Use when fine-tuning LLMs, training custom models, or adapting foundation models for specific tasks. Invoke for configuring LoRA/QLoRA adapters… |
| fine-tuning-openvla-oft 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Fine-tunes and evaluates OpenVLA-OFT and OpenVLA-OFT+ policies for robot action generation with continuous action heads, LoRA adaptation, and FiLM… |
| fine-tuning-serving-openpi 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Fine-tune and serve Physical Intelligence OpenPI models (pi0, pi0-fast, pi0.5) using JAX or PyTorch backends for robot policy inference across… |
| fine-tuning-with-trl 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization… |
| finetuning 86/100 · B | microsoft/azure-skills | Fine-tune models on Microsoft Foundry using SFT (supervised), DPO (preference), or RFT (reinforcement with graders). Covers dataset preparation… |
| gptq 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4×… |
| huggingface-llm-trainer 72/100 · C | huggingface/skills | Train or fine-tune language and vision models using TRL (Transformer Reinforcement Learning) or Unsloth with Hugging Face Jobs infrastructure… |
| huggingface-tokenizers 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram… |
| huggingface-vision-trainer 72/100 · C | huggingface/skills | Trains and fine-tunes vision models for object detection (D-FINE, RT-DETR v2, DETR, YOLOS), image classification (timm models — MobileNetV3… |
| i4h-workflow-finetune 82/100 · B | NVIDIA/skills | Fine-tune a manifest-backed GR00T or openpi remote Task on compatible LeRobot data. Use for training; do not use for inference-only Tasks or… |
| i4h-workflow-train-rl 76/100 · C | NVIDIA/skills | Use when training, evaluating, or exporting Workflow policies with online RSL-RL or RLinf, including RL checkpoint and Workflow handoff. |
| implementing-llms-litgpt 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). Use when need clean… |
| launch-nemo-rl 72/100 · C | NVIDIA/skills | Playbook for launching, monitoring, stopping, and debugging NeMo-RL recipes on a Kubernetes cluster via the nrl-k8s CLI. Covers ephemeral vs… |
| llama-factory 64/100 · D | Orchestra-Research/AI-Research-SKILLs | Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimodal support |
| long-context 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents… |
| mcore-run-on-slurm 70/100 · C | NVIDIA/skills | How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for… |
| microsoft-foundry 79/100 · C | microsoft/azure-skills | Build, deploy, evaluate, optimize, fine-tune, and manage Microsoft Foundry agents, models, and resources end to end. USE FOR: foundry, azd ai agent… |
| miles-rl-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Provides guidance for enterprise-grade RL training using miles, a production-ready fork of slime. Use when training large MoE models with FP8/INT4… |
| moe-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction… |
| nanogpt 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej… |
| nemo-automodel-launcher-config 68/100 · D | NVIDIA/skills | Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution. |
| nemo-automodel-model-onboarding 70/100 · C | NVIDIA/skills | Guide for onboarding new model architectures into NeMo AutoModel, including architecture discovery, implementation patterns, registration, and… |
| nemo-automodel-recipe-development 68/100 · D | NVIDIA/skills | Create and modify NeMo AutoModel training and evaluation recipes, including YAML structure, builders, and execution flow. |
| nemo-mbridge-mlm-bridge-training 68/100 · D | NVIDIA/skills | Run Megatron-LM (MLM) and Megatron Bridge training with mock or real data. Covers correlation testing, available recipes, and multi-GPU examples. |
| nemo-mbridge-multi-node-slurm 73/100 · C | NVIDIA/skills | Convert single-node scripts to multi-node Slurm sbatch jobs and debug common multi-node failures. Covers srun-native vs uv run torch.distributed… |
| nemo-mbridge-perf-activation-recompute 78/100 · C | NVIDIA/skills | Validate and use selective and full activation recompute in Megatron Bridge to reduce GPU memory usage at the cost of extra compute. Use for… |
| nemo-mbridge-perf-cpu-offloading 70/100 · C | NVIDIA/skills | Validate and use CPU offloading in Megatron Bridge, including layer-level activation offloading and fractional optimizer state offloading with… |
| nemo-mbridge-perf-cuda-graphs 70/100 · C | NVIDIA/skills | Validate and use CUDA graph capture in Megatron Bridge, including local full-iteration graphs and Transformer Engine scoped graphs for attention… |
| nemo-mbridge-perf-expert-parallel-overlap 70/100 · C | NVIDIA/skills | Validate and use MoE expert-parallel communication overlap in Megatron-Bridge, including overlap_moe_expert_parallel_comm, delay_wgrad_compute, and… |
| nemo-mbridge-perf-hierarchical-context-parallel 68/100 · D | NVIDIA/skills | Operational guide for enabling hierarchical context parallelism in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification. |
| nemo-mbridge-perf-megatron-fsdp 65/100 · D | NVIDIA/skills | Operational guide for enabling Megatron FSDP in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification. |
| nemo-mbridge-perf-memory-tuning 70/100 · C | NVIDIA/skills | Techniques for reducing peak GPU memory in Megatron Bridge — expandable segments, PEFT + SP input re-gather, parallelism resizing, activation… |
| nemo-mbridge-perf-moe-comm-overlap 68/100 · D | NVIDIA/skills | MoE expert-parallel communication overlap in Megatron Bridge. Covers dispatch/combine overlap, flex dispatcher backends, and expert wgrad scheduling. |
| nemo-mbridge-perf-moe-dispatcher-selection 70/100 · C | NVIDIA/skills | Choose the right MoE token dispatcher (`alltoall`, DeepEP, or HybridEP) for the hardware, EP degree, and optimization stage. Summarizes patterns… |
| nemo-mbridge-perf-moe-hardware-configs 70/100 · C | NVIDIA/skills | Representative, point-in-time MoE training playbooks by hardware and model family. Use them as candidate seeds, then revalidate the exact runtime… |
| nemo-mbridge-perf-moe-long-context 70/100 · C | NVIDIA/skills | Long-context MoE training guidance for Megatron Bridge. Covers CP sizing, selective recompute, dispatcher choices, and practical patterns from DSV3… |
| nemo-mbridge-perf-moe-optimization-workflow 70/100 · C | NVIDIA/skills | Evidence-gated workflow for MoE performance optimization in Megatron Bridge. Covers measurement contracts, the Three Walls framework, parallel… |
| nemo-mbridge-perf-moe-vlm-training 70/100 · C | NVIDIA/skills | Practical guidance for training MoE VLMs in Megatron Bridge. Compares FSDP and 3D-parallel approaches, using rounded lessons from Qwen3-VL… |
| nemo-mbridge-perf-parallelism-strategies 70/100 · C | NVIDIA/skills | Operational guide for choosing and combining parallelism strategies in Megatron Bridge, including sizing rules, hardware topology mapping, and… |
| nemo-mbridge-perf-sequence-packing 70/100 · C | NVIDIA/skills | Validate and use packed sequences and long-context training in Megatron-Bridge, including offline LLM packing, collate-time VLM packing, Energon… |
| nemo-mbridge-perf-tp-dp-comm-overlap 67/100 · D | NVIDIA/skills | Operational guide for enabling TP, DP, and PP communication overlap in Megatron-Bridge, including config knobs, code anchors, pitfalls, and… |
| nemo-mbridge-recipe-recommender 81/100 · B | NVIDIA/skills | Recommend and customize Megatron Bridge library and benchmark recipes for a user's model, GPU count, hardware, sequence length, and… |
| nemo-mbridge-resiliency 68/100 · D | NVIDIA/skills | Resiliency features in Megatron Bridge including fault tolerance, straggler detection, in-process restart, preemption, and re-run state machine. |
| nemo-rl-auto-research 85/100 · B | NVIDIA/skills | Autonomous NeMo-RL research agent workflow for directed hypothesis testing and open-ended discovery. Guides agents through the full experiment… |
| nemotron-asr-finetune 78/100 · C | NVIDIA/skills | Orchestration skill for NVIDIA Nemotron Speech (Riva) / NeMo ASR domain and language adaptation. Given a goal like "improve/fine-tune ASR for my… |
| nemotron-customize 83/100 · B | NVIDIA/skills | Plan, configure, and chain repo-native Nemotron customization steps into single-step or multi-step pipelines: curation, translation, SFT/PEFT… |
| nv-generate-mr-brain-finetune 78/100 · C | NVIDIA/skills | Used for finetuning NV-Generate-CTMR MR-Brain v1 for T1, T2, FLAIR, SWI, or MRA data from a NIfTI datalist. Not for clinical or production data… |
| nv-generate-vae-finetune 76/100 · C | NVIDIA/skills | Used for finetuning the NV-Generate-CTMR MAISI VAE from CT/MRI NIfTI datalists. Not for clinical or production data approval. |
| nv-segment-ct-finetune 81/100 · B | NVIDIA/skills | Runs standard or fixed-channel softmax finetuning of NV-Segment-CT VISTA3D on CT NIfTI image/label datasets, with optional MONAI-native MLflow… |
| openrlhf-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM… |
| peft-fine-tuning 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory… |
| pytorch-fsdp2 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use… |
| pytorch-lightning 81/100 · B | K-Dense-AI/claude-scientific-skills | Deep learning framework (PyTorch Lightning / lightning package). Organize PyTorch code into LightningModules, configure Trainers for multi-GPU/TPU… |
| pytorch-lightning 77/100 · C | Orchestra-Research/AI-Research-SKILLs | High-level PyTorch framework with Trainer class, automatic distributed training (DDP/FSDP/DeepSpeed), callbacks system, and minimal boilerplate… |
| ray-train 77/100 · C | Orchestra-Research/AI-Research-SKILLs | Distributed training orchestration across clusters. Scales PyTorch/TensorFlow/HuggingFace from laptop to 1000s of nodes. Built-in hyperparameter… |
| simpo-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No… |
| slime-rl-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data… |
| stable-baselines3 81/100 · B | K-Dense-AI/claude-scientific-skills | Trains and evaluates single-agent reinforcement learning with Stable Baselines3 (PPO, SAC, DQN, TD3, DDPG, A2C), Gymnasium custom environments… |
| tao-finetune-anomalygennext 91/100 · A | NVIDIA/skills | Validate a user AnomalyGenNext dataset and recipe, freeze its validation set, and prepare a canonical texture fine-tuning recipe. Use when… |
| tao-finetune-clip 89/100 · B | NVIDIA/skills | CLIP vision-language model for image-text retrieval, zero-shot classification, embedding extraction, ONNX export, and TensorRT deployment. Use when… |
| tao-finetune-cosmos-embed 88/100 · B | NVIDIA/skills | Cosmos-Embed1 video-text embedding for text-to-video retrieval, video-to-video search, semantic deduplication, and fine-tuning. Use when the user… |
| tao-finetune-cosmos-reason 79/100 · C | NVIDIA/skills | Shared Cosmos3 frontend that explicitly routes Cosmos Framework and Cosmos-RL, validates runtime model/video-dataset/SLURM inputs, consumes an SQSH… |
| tao-finetune-huggingface-model 85/100 · B | NVIDIA/skills | Fine-tune any HuggingFace CV / VLM / LLM model on local NVIDIA GPUs inside an NGC PyTorch container when no dedicated TAO model skill matches. Use… |
| tao-finetune-kumo-anomaly 88/100 · B | NVIDIA/skills | Kumo-Anomaly — diffusion-based anomaly detection and fine-tuning for multivariate time series. Use when the user asks to "fine-tune Kumo-Anomaly"… |
| tao-finetune-kumo-forecast 88/100 · B | NVIDIA/skills | Kumo-Forecast — transformer-based multivariate time series forecasting with DARR (context-enhanced kNN retrieval), interpretability, and… |
| tao-finetune-video-clip 88/100 · B | NVIDIA/skills | InternVideo2-CLIP L14 (TAO video_clip) for video-text retrieval, zero-shot classification, embedding extraction, LoRA fine-tuning, ONNX export, and… |
| tao-generate-anomalies 88/100 · B | NVIDIA/skills | Full PAIDF AnomalyGen pipeline — fine-tune on a new anomaly dataset, generate synthetic anomaly images (SDG), evaluate quality (nn_score), and… |
| tao-run-automl 90/100 · A | NVIDIA/skills | Run container-backed AutoML / hyperparameter optimization (HPO) for NVIDIA TAO networks using AutoMLRunner. Handles algorithm selection (bayesian… |
| tao-run-automl-deft-pipeline 84/100 · B | NVIDIA/skills | Run the canonical NVIDIA AOI three-phase training pipeline — Phase 1 AutoML baseline (HPO), Phase 2 DEFT loop (RCA → SDG → mining → plain-train… |
| tao-run-deft-aoi-cosmos3 92/100 · A | NVIDIA/skills | Run the disk-backed DEFT AOI improvement loop for NVIDIA Cosmos Reason 3 / Cosmos3 models, using Nano by default and Edge or Super when explicitly… |
| tao-run-on-kubernetes 87/100 · B | NVIDIA/skills | Kubernetes execution platform — submits TAO container jobs as k8s Jobs with NVIDIA GPU scheduling; single-pod for one node, Indexed Jobs for… |
| tao-train-action-recognition 88/100 · B | NVIDIA/skills | Action recognition from video sequences. Supports RGB, optical flow, and joint (multi-stream) input types for classifying temporal actions in video… |
| tao-train-bevfusion 88/100 · B | NVIDIA/skills | BEVFusion for multi-sensor 3D object detection. Fuses LiDAR point clouds and camera images in bird's-eye-view (BEV) space, used in autonomous… |
| tao-train-centerpose 88/100 · B | NVIDIA/skills | CenterPose for keypoint / pose estimation. Detects object centers and regresses keypoint locations for 6-DoF object pose estimation. Use when… |
| tao-train-codetr 88/100 · B | NVIDIA/skills | Co-DETR (CoDINO) for object detection. A DETR-family detector with collaborative hybrid assignment — auxiliary one-to-many heads supervise the… |
| tao-train-deformable-detr 88/100 · B | NVIDIA/skills | Deformable DETR for 2D object detection. Uses deformable attention for efficient multi-scale feature processing, lighter than DINO with competitive… |
| tao-train-depth-anything-v2 88/100 · B | NVIDIA/skills | Monocular depth estimation using Metric Depth Anything v2 or Relative Depth Anything architectures. Predicts per-pixel depth from single RGB images… |
| tao-train-dino 88/100 · B | NVIDIA/skills | DINO (DETR with Improved DeNoising Anchor Boxes) for 2D object detection. Transformer-based detector with denoising training, multi-scale features… |
| tao-train-dinov3 80/100 · B | NVIDIA/skills | DINOv3 continual self-supervised pre-training. Domain-adapts public DINOv3 ViT backbones on unlabeled images via teacher-student self-distillation… |
| tao-train-fast-foundation-stereo 88/100 · B | NVIDIA/skills | Real-time stereo depth estimation using FastFoundationStereo (FFS), the distilled bp2 commercial variant of FoundationStereo. Predicts disparity… |
| tao-train-foundation-stereo 88/100 · B | NVIDIA/skills | Stereo depth estimation using FoundationStereo. Predicts disparity maps from stereo image pairs for 3D reconstruction. Use when training… |
| tao-train-grounding-dino 88/100 · B | NVIDIA/skills | Grounding DINO for open-set object detection. Combines DINO-style detection with a BERT text encoder for language-guided detection — detects objects… |
| tao-train-image-classification 88/100 · B | NVIDIA/skills | PyTorch-based TAO image classification. Supports a wide range of backbones (FAN, EfficientNet, ResNet, etc.) with distillation and quantization for… |
Show the other 27 model training skills
| Skill | Repository | What it does |
|---|---|---|
| tao-train-mask-auto-encoder 88/100 · B | NVIDIA/skills | Masked Auto-Encoder (MAE) for self-supervised pretraining and fine-tuning. Masks random patches and reconstructs them to learn visual… |
| tao-train-mask-auto-label 88/100 · B | NVIDIA/skills | MAL (Mask Auto-Label) for weakly-supervised segmentation. Produces segmentation masks from minimal annotations (point or box annotations) using a… |
| tao-train-mask-grounding-dino 88/100 · B | NVIDIA/skills | Mask Grounding DINO for grounded instance segmentation. Extends Grounding DINO with a mask-prediction head for open-set segmentation guided by text… |
| tao-train-mask2former 88/100 · B | NVIDIA/skills | Mask2Former for universal image segmentation (panoptic, instance, and semantic). Transformer-based with masked attention for high-quality… |
| tao-train-metric-learning-recognition 88/100 · B | NVIDIA/skills | Metric-learning recognition (ml-recog) for fine-grained visual recognition. Learns embeddings for retrieval-based matching (e.g., retail product… |
| tao-train-nvdinov2 88/100 · B | NVIDIA/skills | NVDINOv2 for self-supervised visual representation learning. Trains vision transformers via self-distillation (teacher-student) without labels and… |
| tao-train-nvpanoptix3d 88/100 · B | NVIDIA/skills | NVPanoptix3D for panoptic 3D scene reconstruction from posed RGB images. Produces 3D panoptic segmentation (semantic, instance, and panoptic masks)… |
| tao-train-ocdnet 88/100 · B | NVIDIA/skills | OCDNet for scene text detection. Detects arbitrary-oriented text regions in natural images using a differentiable binarization approach. Use when… |
| tao-train-ocrnet 88/100 · B | NVIDIA/skills | OCRNet for scene text recognition. Recognizes text content from cropped text-region images and supports CTC and attention-based decoders. Use when… |
| tao-train-oneformer 88/100 · B | NVIDIA/skills | OneFormer for universal image segmentation. Unifies panoptic, instance, and semantic segmentation with a single architecture using task-conditioned… |
| tao-train-optical-inspection 90/100 · A | NVIDIA/skills | Optical Inspection for defect detection using Siamese networks. Compares image pairs to detect manufacturing defects, anomalies, or quality issues… |
| tao-train-pointpillars 88/100 · B | NVIDIA/skills | PointPillars for 3D object detection from LiDAR point clouds. Encodes point clouds into a pseudo-image via a pillar-based representation, then… |
| tao-train-pose-classification 88/100 · B | NVIDIA/skills | Pose classification using ST-GCN (Spatial Temporal Graph Convolutional Network). Classifies skeleton sequences into action categories from… |
| tao-train-reid 88/100 · B | NVIDIA/skills | Person re-identification (ReID). Learns discriminative embeddings to match the same person across different camera views, based on metric learning… |
| tao-train-rtdetr 88/100 · B | NVIDIA/skills | RT-DETR (Real-Time DEtection TRansformer) for 2D object detection. Designed for real-time inference with competitive accuracy and supports… |
| tao-train-segformer 88/100 · B | NVIDIA/skills | SegFormer for semantic segmentation. Lightweight transformer-based architecture with hierarchical feature extraction, efficient for real-time… |
| tao-train-single-step 88/100 · B | NVIDIA/skills | Standard single-step train/eval/export workflow for any TAO model. Use when training a TAO model on a dataset without iterative data augmentation… |
| tao-train-sparse4d 87/100 · B | NVIDIA/skills | Prepare calibrated multi-camera data and train, evaluate, run inference, export, or quantize TAO Sparse4D 3D detection and tracking models. Includes… |
| tao-train-visual-changenet 90/100 · A | NVIDIA/skills | Visual ChangeNet for binary image classification and segmentation in AOI defect detection. Use when training, evaluating, exporting, or running… |
| torchdrug 87/100 · B | K-Dense-AI/claude-scientific-skills | Builds and troubleshoots TorchDrug 0.2.1 workflows for molecular graphs, property prediction, self-supervised pretraining, molecule generation… |
| torchforge-rl-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL… |
| training-llms-megatron 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B… |
| transformers 87/100 · B | K-Dense-AI/claude-scientific-skills | Hugging Face Transformers for loading Hub models, running pipeline inference, text generation, and Trainer fine-tuning on NLP, vision, audio, and… |
| trl-training 77/100 · C | huggingface/skills | Post-train LLMs with TRL (Transformers Reinforcement Learning) — SFT, DPO, GRPO, KTO, and reward-model training. Use when writing or debugging… |
| unsloth 64/100 · D | Orchestra-Research/AI-Research-SKILLs | Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization |
| verl-rl-training 69/100 · D | Orchestra-Research/AI-Research-SKILLs | Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL… |
| waypoint-bio 79/100 · C | K-Dense-AI/claude-scientific-skills | Supports work with Outpost Bio's open microbiome foundation models - the Waypoint checkpoints (Waypoint-6m, Waypoint-45m, Waypoint-170m), the Atlas… |
How to choose model training skills
Pick the training library you will run, then the skill for it. Hugging Face (TRL, llm-trainer, vision-trainer), Orchestra-Research (Axolotl, Unsloth, PEFT, DeepSpeed) and NVIDIA (NeMo, Megatron Bridge, TAO) each cover the same ground with their own tooling.
Most of these skills expect GPUs. Check the bundled scripts list and SKILL.md for the container, driver and memory requirements, and for managed tuning, the cloud credentials listed under Before you install.
Keep evaluation in the same plan. Pair a training skill with an eval skill so each run reports a score you can compare, and keep a held-out set the training data never touches.
Related topics
Machine Learning 29Inference and Model Serving 39Bioinformatics and Genomics 59Chemistry and Drug Discovery 28Google Cloud 146Kubernetes 72RAG and Retrieval 32Transcription and Speech to Text 10
Find your next useful skill
Explore Rising Skills with quality scores and source reports, or search from inside your agent.
Get Rising Skills by email · Read Rising Skills · Install the finder · State of Agent Skills