Data Science & ML Claudeスキル

OpenRLHF Model Training

Deploys and manages high-performance RLHF training pipelines for large-scale language models using Ray and vLLM acceleration.

NVIDIA TensorRT-LLM Optimization

Optimizes Large Language Model inference for maximum throughput and ultra-low latency on NVIDIA GPUs.

SAELens Mechanistic Interpretability

Decomposes neural network activations into interpretable, sparse features using SAELens for deep mechanistic interpretability research.

BLIP-2 Vision-Language

Implements state-of-the-art vision-language pre-training to enable high-quality image captioning and visual question answering within AI workflows.

SentencePiece Tokenization

Implements language-independent subword tokenization using BPE and Unigram algorithms for robust NLP model training and inference.

HuggingFace Accelerate Distributed Training

Simplifies PyTorch distributed training across multiple GPUs, TPUs, and nodes with minimal code changes and a unified API.

Sentence Transformers Embeddings

Generates high-quality sentence, text, and image embeddings for RAG, semantic search, and clustering using state-of-the-art transformer models.

Ray Data Scalable Processing

Facilitates high-performance distributed data processing and streaming for large-scale machine learning workloads.

Weights & Biases MLOps

Integrates Weights & Biases into your workflow to track machine learning experiments, visualize training metrics, and manage model artifacts in real-time.

torchforge RL Training

Implements PyTorch-native agentic reinforcement learning workflows using Meta's torchforge library for scalable algorithm experimentation.

Mamba Architecture Guide

Implements and optimizes Selective State Space Models (SSM) for high-performance sequence modeling and long-context AI applications.

Pyvene Causal Interventions

Facilitates causal interventions on PyTorch models using a declarative framework for mechanistic interpretability experiments.

DeepSpeed Distributed Training

Optimizes large-scale model training using DeepSpeed configurations, ZeRO optimization stages, and high-performance I/O management.

MoE Training: Mixture of Experts

Trains and optimizes Mixture of Experts (MoE) models to scale AI capacity with significantly reduced compute costs.

RWKV Model Architecture

Implements and manages RWKV architectures for efficient, linear-time AI inference and long-context processing.

LM Evaluation Harness

Evaluates Large Language Models across 60+ academic benchmarks using standardized prompts and metrics for reproducible research.

Whisper Speech Recognition

Transcribes and translates audio across 99 languages using OpenAI's robust general-purpose speech recognition models.

GPTQ Model Quantization

Compresses large language models to 4-bit precision to enable high-speed inference and deployment on consumer-grade hardware.

Axolotl LLM Fine-Tuning

Streamlines the fine-tuning of large language models using Axolotl through expert YAML configuration and advanced training pattern guidance.

AudioCraft Audio Generation

Generates high-quality music and sound effects from text descriptions using Meta's AudioCraft library.

PEFT LLM Fine-Tuning

Optimizes large language model fine-tuning using LoRA, QLoRA, and other parameter-efficient methods to significantly reduce memory and hardware requirements.

SimPO LLM Alignment

Implements Simple Preference Optimization to align Large Language Models without requiring a reference model.

vLLM Inference Serving

Serves large language models with high throughput and low latency using PagedAttention and continuous batching.

LangChain Framework Integration

Builds sophisticated LLM applications using agents, chains, and Retrieval-Augmented Generation (RAG) with a unified interface.

nnsight Remote Interpretability

Interprets and manipulates neural network internals across local and remote models using the nnsight library and NDIF execution.

AWQ LLM Quantization & Optimization

Optimizes Large Language Models using 4-bit activation-aware weight quantization to achieve 3x faster inference with minimal accuracy loss.

BigCode Evaluation Harness

Evaluates AI code generation models using industry-standard benchmarks and pass@k metrics.

PyTorch Lightning Training

Standardizes and accelerates PyTorch model training with built-in support for distributed computing, logging, and engineering best practices.

Knowledge Distillation for LLMs

Compresses large language models into efficient student models while retaining performance through advanced teacher-student transfer techniques.

HuggingFace Tokenizers

Provides high-performance, Rust-based tokenization tools for building and training NLP models with support for BPE, WordPiece, and Unigram algorithms.