Claude Skills by VectorSpaceLab

"Use for Ultralytics track mode, tracker YAML selection, persistent

"Plan and troubleshoot Ultralytics YOLO training, validation,

"Routes UltralyticsPro workflows for YOLO and RT-DETR training,

"Routes UltralyticsPro prediction workflows for YOLO image

"Routes UltralyticsPro training workflows for YOLO and RT-DETR

"Use the umap-learn package for UMAP dimensionality reduction,

"Covers AlignedUMAP for related slices and fitted UMAP model

"Use base umap.UMAP for unsupervised embedding, transforms, input

"Use ParametricUMAP for TensorFlow/Keras-backed neural embeddings,

"Diagnose optional umap.plot static, datashaded, connectivity,

"Supervised, semi-supervised, densMAP, clustering, and outlier

"Routes UNet++ work across the official PyTorch nnU-Net stack and

"Guides the legacy Keras segmentation-models stack for UNet++,

"Guides the official PyTorch nnU-Net workflow for UNet++ training,

"Operate the UniAD repository for planning-oriented autonomous

"Explain UniAD config variants, plugin registration, public model

"Prepare and validate UniAD nuScenes, CAN bus, map, info-PKL, and

"Build and troubleshoot UniAD distributed and SLURM train/eval runs

"Interpret UniAD result artifacts and build

"Use the Microsoft UniLM umbrella repository for language

"Plan and sanity-check UniLM architecture research,

"Use UniLM E5 and SimLM for embedding evaluation, dense retrieval,

"Plan and troubleshoot UniLM language sequence-to-sequence

"Use UniLM Kosmos, TextDiffuser, WavLM, BEATs, SpeechT5, SpeechLM,

"Use UniLM vision, vision-language, OCR, webpage, and document-AI

"Use Darts for Python time-series construction, preprocessing,

"Use Darts anomaly scorers, detectors, aggregators, and forecasting

"Use Darts preprocessing transformers, Pipeline, missing-value

"Use Darts deterministic/probabilistic metrics, reductions,

"Use Darts core forecasting workflows, model selection, covariates,

"Use Darts TimeSeries constructors, shapes, static covariates,

"Use Darts PyTorch forecasting models and foundation wrappers with

"Routes UnrealCV tasks across the Python client, plugin build, and

"Routes UnrealCV command-doc, public API snapshot, and

"Routes UnrealCV plugin build, install, and packaging tasks for

"Routes UnrealCV Python client tasks for connecting to servers,

"Use Unsloth Core, CLI, Studio, and export workflows for fast LLM

"Use the Unsloth Typer CLI for training, inference, chat, export,

"Plan and validate code-first Unsloth Core model loading, LoRA/full

"Save and export Unsloth checkpoints as LoRA adapters, merged

"Launch, secure, connect, operate, and troubleshoot Unsloth Studio

"Use Unstract to operate its document-extraction APIs, hosted MCP

"Use backend-platform for Unstract's Django/DRF APIs, hosted MCP

"Use frontend for Unstract's React/Vite/Bun browser app, route

"Use Unstract platform-deployment to launch, inspect, and

"Use sdk-and-tools for Unstract's shared Python packages, tool

"Use testing-rig for Unstract's manifest-driven unit, integration,

"Use workers for Unstract's Celery and Postgres-backed queue

"Use the Unstructured Python package to partition documents into

"Configure and validate Unstructured post-partition chunking for