Skill UI

torchforge is Meta's advanced PyTorch-native library designed for agentic Reinforcement Learning (RL). It fundamentally separates core RL algorithms from complex distributed infrastructure concerns. This allows researchers to rapidly prototype and experiment with novel RL methods (like GRPO) using clean, PyTorch-native abstractions, while automatically handling large-scale tasks like distributed training across thousands of GPUs using Monarch and TorchTitan.

PufferLib is a high-performance library designed for advanced reinforcement learning simulations and training. It excels in parallel environment simulation, offering optimized vectorization and native support for multi-agent systems. Utilizing algorithms like PuffeRL (optimized PPO+LSTM), it achieves millions of steps per second, significantly accelerating research and deployment across single-agent and multi-agent environments.

Stable Baselines3 is a robust, PyTorch-based library providing production-ready implementations of core reinforcement learning algorithms (PPO, SAC, DQN, etc.). It facilitates the training of single-agent RL models, supports the creation of custom Gymnasium environments, and handles advanced parallel training via vectorized environments. It is ideal for rapid prototyping and rigorous RL experiments.

PufferLib is a high-performance, scalable RL framework designed for massive parallel training and vectorized simulation. Featuring PuffeRL (optimized PPO+LSTM) and the PufferEnv API, it supports multi-agent systems, custom environment creation, and achieves industry-leading speeds (up to millions of steps/second). Ideal for advanced research and large-scale AI experimentation.

Stable Baselines3 is a robust, PyTorch-based library providing reliable implementations of state-of-the-art reinforcement learning algorithms (PPO, SAC, DQN, etc.). It is designed for standard RL experiments, quick prototyping, and developing complex single-agent systems. The library supports training agents, creating custom Gymnasium environments, and running vectorized simulations for accelerated learning.

A comprehensive, AI-powered suite designed to generate high-quality, platform-optimized content for LinkedIn. It creates full posts, multi-slide carousels, long-form newsletters, and 30-day content calendars. The tool integrates niche-specific SEO rules and features a reinforcement-learning memory system, ensuring that your content style and performance improve with every piece of feedback.

This skill provides expert capabilities for training and fine-tuning transformer language models using the TRL (Transformers Reinforcement Learning) library. It supports state-of-the-art post-training methods, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and Reward Model training. Use this skill to align and customize foundation models for advanced tasks via CLI commands.