Tag
Explore by tags
Gives developers low-level primitives for building stateful single-agent, multi-agent, and graph-based control flows, with built-in human-in-the-loop checkpoints, persistent cross-session memory, and token-level streaming.
A PyTorch-native, hardware-agnostic stack for robot learning: data collection, training, and deployment across 11+ robots, from SO100 to Unitree G1. Includes imitation, RL, and vision-language-action policies (ACT, Diffusion, Pi0, SmolVLA).
Runs AI-generated code in isolated, elastic sandboxes with SDK, API, and CLI access for agent workflows that need stateful execution and environment control.
Automates online monetization workflows—generating and scheduling YouTube Shorts, posting to X (Twitter), running affiliate campaigns, and outreach. Modular provider-based design (TTS, LLM hooks, CRON scheduler) and configurable pipelines; legal/ToS risks mean use with caution.
Runs coding agents and automations from a self-hosted developer control center, with local, remote, cloud, and ACP-compatible backends for managed engineering workflows.
Framework for building offensive and defensive security agents that run real pentests autonomously. Uses a ReACT loop over 300+ models (OpenAI, Anthropic, DeepSeek, local Ollama) with built-in recon, exploitation, and privilege-escalation tools.
A minimal GPU written in under 15 SystemVerilog files to teach how GPUs execute parallel kernels from the ground up. Includes an 11-instruction ISA, multiple cores with ALUs and load-store units, a fetch-decode-execute pipeline, and matrix kernels.
Provides a cleaned, deduplicated English web corpus optimized for LLM pretraining—over 15T tokens aggregated from CommonCrawl with per-dump snapshots and smaller sampled configs (10B/100B/350B). Includes the datatrove processing pipeline, MinHash deduplication, and an ODC-By v1.0 license; suited for large-scale model training and ablation studies but not specialized for code.
Chains pre-trained AI weather and climate models like GraphCast, Pangu, and FourCastNet into composable inference pipelines. Swap prognostic or diagnostic components, plug in reanalysis sources, and add ensemble perturbations or in-loop metrics.
React components for building LLM chat and agent interfaces: message bubbles, prompt sets, conversation lists, and sender inputs under a RICH interaction paradigm, plus a streaming Markdown renderer and hooks for wiring UI to model data streams.
Splits autonomous R&D into two cooperating agents: one proposes hypotheses, the other writes and tests code — iterating on quant-finance factors, Kaggle pipelines, and model research. Hits a ~30% medal rate on MLE-Bench, nearly double AIDE's.
A community speedrun to train a 124M GPT as fast as possible on 8 H100s, all chasing a fixed 3.28 FineWeb loss. Successive records cut the run from llm.c's 45 minutes to under 1.4, mostly via the new Muon optimizer rather than more hardware.