Skill UI

Guides engineers through DeepSpeed distributed training and NVMe I/O optimizations, covering ZeRO, pipeline parallelism, mixed precision, and DeepNVMe handles for synchronous and asynchronous tensor storage.

Provides detailed evaluation of machine learning models using plugin metrics to assess accuracy, precision, recall, F1, and other key indicators, helping users compare models and validate performance before deployment.

Provides calibration-free weight quantization for LLMs down to 1-bit precision with multiple optimized backends, ideal for fast deployments with vLLM/HuggingFace and LoRA/PEFT fine-tuning without collecting calibration data.

HuggingFace Accelerate adds distributed support to any PyTorch script with just four lines, providing a unified API for DDP, DeepSpeed, FSDP, and Megatron, plus automatic device placement, mixed precision, and simple launch/config commands for fast prototyping.

Guides production-grade miles RL training for massive MoE models, covering FP8/INT4 low precision, train-inference alignment, and speculative rollout workflows to maximize throughput in enterprise deployments.

Automates mixed precision ML training workflows by surfacing best practices, generating runnable code/configurations, and validating setups across data prep, model tuning, and experiment tracking.

Flash Attention optimizer delivers 2-4x speedup and 10-20x memory savings for transformer attention, ideal when training or running long-context models that hit GPU memory or latency limits; supports PyTorch native scaled_dot_product_attention, the flash-attn library, H100 FP8 precision, and sliding-window attention.

Guides adding PyTorch FSDP2’s fully_shard integration to training loops, covering correct distributed init, bottom-up sharding, mixed precision/offload policies, DTensor-aware optimizers, and distributed checkpointing so oversized models or DeviceMesh sharding goals work reliably.

PyTorch Lightning is a high-level framework built on PyTorch that significantly reduces boilerplate code in deep learning training loops. It automatically handles device management, distributed training (DDP, FSDP, DeepSpeed), mixed precision, and advanced callbacks. It allows researchers to scale models from local machines to supercomputers with minimal code changes, enabling clean, scalable, and best-practice deep learning implementations.

Optimizes large language model inference on NVIDIA GPUs with TensorRT, delivering 10-100× faster throughput, quantized precision (FP8/INT4), multi-GPU scaling, and serving-ready tooling for production deployments.

Provide REST API access to ClinPGx, the successor to PharmGKB, for querying gene-drug interactions, CPIC/DPWG/FDA guidelines, allele functions, and evidence that supports precision dosing and pharmacogenomics-driven clinical decisions.

This skill provides comprehensive techniques for optimizing prompts used with any Large Language Model (LLM). It teaches proven methodologies—including Role definition (R), Context setting (S), Constraints (C), Instructions (I), and Templates (T)—to transform vague inputs into precision-engineered instructions. Users can reliably achieve structured JSON output, implement Chain-of-Thought reasoning, reduce hallucinations, and minimize token usage, ensuring consistent and high-quality results across different AI platforms.