A machine learning software for extracting information from scholarly documents

EAGLE: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders

Text-to-video generation: CogVideoX (2024) and CogVideo (ICLR 2023)

Bring projects, wikis, and teams together with AI. AppFlowy is an AI collaborative workspace where you achieve more without losing control of your data. The best open source alternative to Notion.

🕷️ The pipeline for the OSCAR corpus

Official repository for KoMT-Bench built by LG AI Research

The most accurate natural language detection library for Python, suitable for short text and mixed-language text

Topic Modelling for Humans

Tool to detect the ISO-15924 code of the script used in a text

C source that once compiled can generate a TECkit map from a Unihan_Variants.txt file to map traditional Chinese characters to simplified Chinese characters

Code for the paper: Don't Settle for Average, Go for the Max: Fuzzy Sets and Max-Pooled Word Vectors, ICLR 2019.

Data and tools for generating and inspecting OLMo pre-training data.

Implementations of online merging optimizers proposed by Online Merging Optimizers for Boosting Rewards and Mitigating Tax in Alignment

Editing Models with Task Arithmetic

GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection

A library for easily merging multiple LLM experts, and efficiently train the merged LLM.

Code associated with the Don't Stop Pretraining ACL 2020 paper

Tools for merging pretrained large language models.

OmniCorpus: A Unified Multimodal Corpus of 10 Billion-Level Images Interleaved with Text

Zstandard - Fast real-time compression algorithm

[SIGGRAPH Asia 2024, Journal Track] ToonCrafter: Generative Cartoon Interpolation

Central place for the engineering/scaling WG: documentation, SLURM scripts and logs, compute environment and data.

Official code for ReLoRA from the paper Stack More Layers Differently: High-Rank Training Through Low-Rank Updates

A modern JavaScript library for handling Hangul characters.

Mediapipe-based library to redact faces from videos and images

Infinity is a high-throughput, low-latency REST API for serving text-embeddings, reranking models and clip