Skill UI

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.

Outlines guarantees valid JSON/XML/code outputs by applying CFG/FSM-constrained token filtering, Pydantic schemas, and grammar-level sampling while supporting local Transformers, llama.cpp, and vLLM backends for fast inference.

Guides training large language models with Volcano Engine's verl library, covering PPO/GRPO workflows, distributed backends, multi-turn rollouts, reward configs, critic-enabled PPO, and production deployments using Ray workers.

This skill provides comprehensive tools for documenting and managing model performance on Hugging Face. It supports multiple methods for adding structured evaluation data, including extracting tables from README content, importing benchmark scores via APIs (like Artificial Analysis), and executing high-performance custom evaluations using advanced backends such as vLLM or lighteval. It ensures that all metadata is structured and compatible with model leaderboards and industry standards.

This skill implements an autonomous, multi-round review loop for refining complex research or codebases. It iteratively reviews content using specified external backends (e.g., Codex, manual review), implements necessary fixes based on feedback, and re-reviews until the positive assessment criteria are met or the maximum round limit is reached. It manages state persistence and allows for configurable difficulty levels, making it ideal for projects requiring rigorous, self-correcting quality assurance.

Get a deep, multi-round, and adversarial critical review of your research work, ideas, or drafts. This skill simulates the perspective of a senior ML reviewer (e.g., NeurIPS/ICML level) using dedicated external backends (Codex or Manual). It is designed to find logical gaps, missing experiments, and narrative weaknesses, providing brutal honesty to elevate your academic contribution.

This skill is designed for running comprehensive evaluations of Hugging Face Hub models on local hardware. It supports multiple evaluation frameworks, including `inspect-ai` and `lighteval`, and provides flexibility in backend selection (e.g., `vllm`, Transformers, or `accelerate`) to optimize for performance and compatibility. Use this for local GPU benchmarking, smoke testing, and backend selection, but note that it does not handle remote job orchestration or result publishing.

Transformers.js allows developers to run state-of-the-art machine learning models directly within JavaScript/TypeScript environments (browser, Node.js). It supports a comprehensive range of tasks including NLP (text generation, summarization, translation), Computer Vision (image classification, object detection), and Audio processing. This capability enables building fully client-side, multimodal AI applications without requiring a dedicated Python backend or server infrastructure.

This skill guides the generation of a complete, production-ready Model Context Protocol (MCP) server using C# and .NET. It covers advanced architectural patterns, including Host builder setup, proper dependency injection, structured logging, and implementing robust functional tools. Ideal for developers building sophisticated AI agents or backend services that require standardized communication and tool invocation capabilities.

This tool audits the integrity of machine learning experiments by checking for common fraud patterns such as fake ground truth, score normalization fraud, and phantom results. It utilizes an external reviewer backend (like Codex or manual review) to perform an adversarial review across all project artifacts, ensuring reported results are trustworthy and methodologically sound.

This skill integrates LangChain 1.0 and LangGraph 1.0 into standard OpenTelemetry (OTEL) backends (e.g., Jaeger, Honeycomb, Tempo). It solves critical observability gaps by ensuring correct prompt content capture, propagating spans across complex subgraphs, and providing five LLM-specific Service Level Objectives (SLOs) like latency and cost. Ideal for multi-tenant, compliance-sensitive, or existing OTEL stack environments where LangSmith is unsuitable.

Generates high-quality, short video clips (3-15s) using the Seedance 2.0 model. This tool accepts structured English prompts, an optional first-frame image, and an optional identity/style reference image for continuity. It supports dual backends (OpenRouter, Volcengine, BytePlus) to maximize accessibility and reliability for cinematic shot generation.