Claude Skills by VectorSpaceLab

"Rank pretrained models with TLLib transferability metrics

"Operate Transfer Learning Library semi-supervised and

"Operate Transfer Learning Library domain-generalization and

"Operate TLLib image/domain translation and style-transfer

"Operate TLLib vision datasets, ImageList formats, transforms,

"Route Transformer Engine installation/build, PyTorch, JAX, and

Install, build, and verify NVIDIA Transformer Engine runtime variants safely.

"Use NVIDIA Transformer Engine JAX and Flax APIs for module

"Operate transformer_engine.pytorch for PyTorch layers, precision

"Operate on the Transformer Lab application monorepo: React web UI,

"Modify and debug Transformer Lab FastAPI backend routers,

"Operate on Transformer Lab Typer CLI, Textual job monitor, CLI

"Operate on the Transformer Lab React TypeScript web UI,

"Operate on Transformer Lab task creation, job launch queues,

"Use and extend Hugging Face Transformers for inference,

"Configure Transformers text/chat generation, GenerationConfig,

"Use Transformers pipelines and AutoClass inference workflows

"Add or modify Transformers models, tokenizers, processors,

"Choose and validate Transformers quantization methods, backend

"Use the Transformers CLI for download, chat, local serving,

"Use Transformers tokenizers, processors, image/audio/video

"Create Transformers Trainer, TrainingArguments, data-collator,

"Guides TransFuser autonomous-driving model training, multimodal

"Plans and diagnoses TransFuser CARLA 0.9.10.1 evaluation, Longest6

"Guides TransFuser CARLA training route and scenario preparation,

"Guides TransFuser imitation-model training, dataset and

"Operate a trained TransFuser model through the CARLA sensor-agent

"Routes Tree of Thoughts workflows for game24, coherent passage

"Routes 5x5 mini crossword workflows for Tree of Thoughts sampling,

"Routes Game of 24 workflows for arithmetic puzzle solving with

"Routes coherent passage generation workflows for the Tree of

"Use the tree-of-thoughts Python package to configure TotAgent

"Configure and inspect tree-of-thoughts TotAgent model adapters,

"Run and adapt tree-of-thoughts DFS/BFS search workflows with

"Route NVIDIA Triton Inference Server tasks across server runtime,

"Build and troubleshoot Triton KServe HTTP/gRPC request payloads

"Use Triton Python packages `tritonserver` and `tritonfrontend` to

"Create, validate, and troubleshoot Triton model repositories and

"Launch and use Triton's OpenAI-compatible LLM frontend over vLLM

"Plan, launch, build, deploy, observe, and troubleshoot NVIDIA

"Use and modify TRL, the Hugging Face Transformers Reinforcement

"Build and debug TRL CLI commands, YAML configs, dataset mixture

"Choose, instantiate, and troubleshoot TRL stable post-training

"Prepare and validate TRL datasets, chat templates, multimodal

"Use TRL experimental trainers and environment-based GRPO

"Modify the TRL repository safely as a contributor or maintainer,

"Choose and diagnose TRL scaling, memory, kernel, PEFT, and vLLM

"Guides trlX LLM post-training, RLHF, Accelerate, configuration,

"Router for trlX's optional NVIDIA NeMo/Megatron backend: when to

"Training, configuration, data, checkpointing, PEFT, and sweep