Download

Skill UI

Browse and discover 5115+ curated skills

All Development Artificial Intelligence Design & Creative Product & Business Data Science Marketing Soft Skills Productivity Engineering Languages

Search Transformer , found 16 results

Default Newest Most Downloaded

Neural Network Builder

building-neural-networks

jeremylongshore/claude-code-plugins-plus-skills

Enables Claude to configure neural network architectures through the neural-network-builder plugin, creating or modifying models and defining layers, activation functions, and training parameters whenever network design or tuning is requested.

GPTQ LLM Quantization

Orchestra-Research/AI-Research-SKILLs

Post-training 4-bit quantization for LLMs that trims memory by 4× with minimal perplexity loss, enabling 70B+ models to run faster on consumer GPUs while integrating with transformers, PEFT, and QLoRA workflows.

Fast NLP Tokenizers

huggingface-tokenizers

Orchestra-Research/AI-Research-SKILLs

Rust-powered HuggingFace tokenizers deliver <20s per GB speed, trainable BPE/WordPiece/Unigram models, alignment tracking, padding, and transformers integration for production NLP pipelines.

Long Context Extensions

Orchestra-Research/AI-Research-SKILLs

Extends transformer models’ context windows using RoPE, YaRN, ALiBi, and interpolation so LLMs can process documents of 32k–128k+ tokens, extrapolate to longer lengths, and deploy efficient positional encodings and bias strategies for fine-tuning or inference.

Mamba Selective State Models

mamba-architecture

Orchestra-Research/AI-Research-SKILLs

Mamba provides selective state-space models with O(n) inference complexity, letting you handle million-token sequences faster than transformers while skipping KV caches and benefiting from a hardware-aware design. Use it for long-context language modeling, streaming applications, and scalable low-memory sequence learners.

Minimalist GPT Training

Orchestra-Research/AI-Research-SKILLs

NanoGPT is a hackable ~300-line PyTorch reproduction of GPT-2 for learning transformer internals; train Shakespeare or OpenWebText on CPUs or multi-GPU setups, experiment with fine-tuning/custom datasets, and prototype quickly.

Clinical AI SDK Patterns

openevidence-sdk-patterns

jeremylongshore/claude-code-plugins-plus-skills

Guides building clinical AI integrations with the OpenEvidence SDK, covering singleton client setup, typed query builders, response transformers, and caching strategies for reliable decision support.

Flash Attention Optimization

optimizing-attention-flash

Orchestra-Research/AI-Research-SKILLs

Speeds up transformer attention with Flash Attention techniques, offering 2-4× throughput gains and 10-20× memory reduction. Ideal for long-context models on PyTorch (native or flash-attn lib), H100 FP8, and sliding window scenarios to resolve GPU memory bottlenecks and improve inference.

Structured Text Generation

Orchestra-Research/AI-Research-SKILLs

Outlines guarantees valid JSON/XML/code generation via CFG-driven FSM filtering, Pydantic schemas, and fast local or API models (Transformers, vLLM, llama.cpp), making structured inference safe and high-performance.

LLM Quantization Toolkit

quantizing-models-bitsandbytes

Orchestra-Research/AI-Research-SKILLs

Quantizes LLMs to 8-bit or 4-bit to cut memory by 50-75%, letting inference and QLoRA fine-tuning run on constrained GPUs while keeping accuracy high; supports Bitsandbytes, NF4/INT8 formats, 8-bit optimizers, and HuggingFace Transformers.

RWKV Hybrid Architecture

rwkv-architecture

Orchestra-Research/AI-Research-SKILLs

RWKV hybrid architecture merges parallel Transformer training with sequential RNN inference to deliver linear-time, infinite-context generation without KV caches, enabling streaming text production and long-document processing on constrained hardware.

Advanced Vision Engineering Guide

senior-computer-vision

alirezarezvani/claude-skills

Guides production-level computer vision engineering with object detection, segmentation, model optimization, and deployment, covering YOLO/Faster R-CNN/DETR, Mask R-CNN/SAM, vision transformers, PyTorch tooling, ONNX/TensorRT pipelines, and deployment best practices.

Language