A collection of Pythonic subword tokenisers and text preprocessing tools.

Ai safety research showing a working bypass to Google's synthid on Nano Banana Pro

Code for Bolmo: Byteifying the Next Generation of Language Models

Open and efficient video and image watermarking

Official implementation of the paper "The Stable Signature Rooting Watermarks in Latent Diffusion Models"

Official implementation of the paper "Watermarking Autoregressive Image Generation" (NeurIPS'25)

Meta Seal is an open-source suite of watermarking models, code, and research developed by Meta to advance the state of content authenticity and attribution across multiple modalities.

InternVL-U is a 4B-parameter unified multimodal model (UMM) that brings multimodal understanding, reasoning, image generation, image editing into a single framework.

Symphony turns project work into isolated, autonomous implementation runs, allowing teams to manage work instead of supervising coding agents.

Google Workspace CLI — one command-line tool for Drive, Gmail, Calendar, Sheets, Docs, Chat, Admin, and more. Dynamically built from Google Discovery Service. Includes AI agent skills.

This repository contains a summary of knowledge cut-off dates for various large language models (LLMs), such as GPT, Claude, Gemini, Llama, and more.

Tiny, Fast, and Deployable anywhere — automate the mundane, unleash your creativity

Fast face detection, pupil/eyes localization and facial landmark points detection library in pure Go.

TabiBERT is a modernized encoder-only model pre-trained on 1 trillion tokens from scratch!

Linear Attention for Efficient Bidirectional Sequence Modeling

Training and evaluation code for the paper "Headless Language Models: Learning without Predicting with Contrastive Weight Tying" (https://arxiv.org/abs/2309.08351)

dLLM: Simple Diffusion Language Modeling

A minimal implementation of DeepMind's Genie world model

Codebase for SentenceSmith

Fast and memory-efficient exact attention

Data augmentation for NLP

👷 Build images with images

Multilingual Pretraining for Pixel Language Models (EMNLP 2025, Official Implementation)

Code for the "Sample-efficient Integration of New Modalities into Large Language Models" paper

This is the official code release for Bayesian Flow Networks.

Qwen-Image is a powerful image generation foundation model capable of complex text rendering and precise image editing.

Local, open-source AI app builder for power users ✨ v0 / Lovable / Replit / Bolt alternative 🌟 Star if you like it!

A dynamic forecasting benchmark for LLMs