Claude Skills by VectorSpaceLab

"Routes Papers-in-100-Lines optimizer, activation, layer,

"Locate paper implementations and produce safe catalog-based run or

"Use PARL reinforcement-learning framework workflows, core

"Select and adapt PARL built-in reinforcement-learning algorithms

"Build and debug PARL core Model, Algorithm, and Agent code, select

"Use PARL environment wrappers, replay buffers, schedulers, CSV

"Build, inspect, and safely use PARL EvoKit, the optional C++

"Operate PARL's optional all-GPU Waymax-RL autonomous-driving

"Guide safe PARL xparl cluster lifecycle, remote_class actors,

"Guide PDEBench scientific-machine-learning benchmark workflows for

"Safely inspect PDEBench dataset metadata and local HDF5 data,

"Route safe, reproducible PDEBench simulation generation, Hydra/JAX

"Route PDEBench forward and inverse model configuration, safe model

"Use PEFT for parameter-efficient fine-tuning adapters, LoRA and

"Core PEFT adapter setup, configuration, target module selection,

"Configure PEFT LoRA, LoRA variants, and quantized-base workflows,

"Configure and troubleshoot PEFT prompt-learning and soft-token

"Use when modifying PEFT itself, preparing a PEFT pull request,

"Save, load, inspect, merge, hotswap, mix, and convert PEFT

"Select, configure, and troubleshoot PEFT's specialized

"Use PEFT with Transformers Trainer, Accelerate, FSDP, DeepSpeed,

"Use PennyLane for quantum circuits, differentiable QNodes,

"Use PennyLane application modules for qchem, mappings, QAOA,

"Build and debug PennyLane QNodes, devices, measurements, shots,

"Configure PennyLane gradients, differentiable interfaces,

"Use PennyLane circuit import/export, dataset loading, debugging

"Use PennyLane operations, templates, observables, op math,

"Work safely in a PennyLane source checkout: focused tests,

"Routes LPIPS image-similarity, BAPPS evaluation, and BAPPS

"Routes BAPPS 2AFC and JND evaluation workflows for LPIPS,

"Routes BAPPS training, fine-tuning, checkpointing, and smoke-test

"Routes LPIPS pairwise comparison, directory comparison, spatial

"Use Petals for distributed large-language-model inference, prompt

"Construct safe Petals benchmark command templates, choose tiny

"Use Petals client APIs for distributed model loading, generation,

"Operate Petals RemoteSequential block sequences, local block

"Operate Petals prompt-tuning, deep prompt-tuning, and

"Operate Petals server and DHT swarm commands for public capacity,

"Routes Petastorm workflows for reading, writing, and converting

"Routes Petastorm dataset creation, schema design, metadata repair,

"Routes Petastorm dataset-reading workflows, including plain Python

"Routes Spark DataFrame conversion workflows that materialize

"Use, select, author, validate, wrap, and integrate PettingZoo

"Create and review PettingZoo AEC and Parallel custom environments,

"Choose PettingZoo environment families, install minimal extras,

"Run and interpret PettingZoo compliance, seed, render, max-cycles,

"Adapt PettingZoo training tutorials for CleanRL, Tianshou, SB3,

"Use PettingZoo AEC and Parallel environments safely, including

"Use PettingZoo conversion wrappers, utility wrappers,

"Routes PFLlib tasks for dataset preparation, federated-learning