Claude Skills by VectorSpaceLab

"Routes LightX2V generation, serving, disaggregation, and

"Routes LightX2V LoRA extraction, LoRA merging, dummy-meta export,

"Routes LightX2V disaggregated deployment workflows for the

"Routes direct LightX2V generation and model-preparation workflows

"Routes LightX2V FastAPI service workflows, task APIs,

"Operate LightFM recommendation models, data/feature matrices,

"Build LightFM Dataset mappings, sparse interactions, weights,

"Evaluate fitted LightFM recommenders and create leakage-safe

"Train, resume, tune, inspect, serialize, score, and troubleshoot

"Maintain LightFM editable builds, compiled extension variants,

"Use LightGlue for local feature matching, extractor selection,

"Choose LightGlue-supported feature extractors and validate feature

"Match two images end-to-end with LightGlue and a selected

"Configure and call LightGlue directly on feature dictionaries and

"Benchmark LightGlue latency and throughput, and create or save

"Route LightLLM serving, deployment, and validation workflows into

"Validate LightLLM with repo-native benchmarks, regressions, static

"Plan LightLLM launch topologies, process sequencing, ports, and

"Explain LightLLM model-family support, backend selection,

"Operate LightLLM HTTP serving, request payloads, streaming, and

"Use LightlySSL for self-supervised computer-vision training, model

"Use LightlySSL CLIs, Hydra overrides, data layouts, YOLO crops,

"Use LightlySSL evaluation utilities and choose safe repository

"Assemble Lightly low-level SSL data, transform, loss, head, and

"Lightly SSL training recipes for PyTorch, Lightning, and gated

"Operate Lightning-Hydra-Template projects: Hydra configs,

"Configure Lightning-Hydra-Template experiments with Hydra

"Customize Lightning-Hydra-Template DataModules, LightningModules,

"Maintain Lightning-Hydra-Template tests, CI, Makefile targets,

"Run and debug Lightning-Hydra-Template training and evaluation

"Build, configure, debug, distribute, and deploy PyTorch Lightning

"Build and troubleshoot LightningCLI and LightningArgumentParser

"Export, optimize, validate, and serve Lightning models for

"Choose, configure, validate, and troubleshoot Lightning Trainer

"Use this sub-skill for Lightning Fabric expert-controlled PyTorch

"Build, convert, debug, and validate core Lightning training

"Use LightRAG as a graph-based RAG framework: embedded Python APIs,

"Operate the LightRAG API server, WebUI, route families, auth,

"Use LightRAG as an embedded Python library: construct LightRAG,

"Work with LightRAG document ingestion internals: parser routing,

"Configure and troubleshoot LightRAG LLM, embedding, VLM,

"Configure and troubleshoot LightRAG storage backends, workspace

"Use LimiX for structured/tabular foundation-model inference,

"Use LimiX benchmark-style classification and regression CLI

"Inspect, choose, generate, and debug LimiX inference configs and

"Use LimiXPredictor for local checkpoint inference on tabular

"Operate LimiX sample-retrieval ensemble inference and Optuna

"Use for LiteLLM Python SDK, AI Gateway proxy, model routing,

"Use when configuring LiteLLM for MCP tools, A2A agents, Claude

"Use when mapping LiteLLM provider/model prefixes, selecting SDK or