Running on CPU Upgrade 192 The Synthetic Data Playbook: Generating Trillions of the Finest Tokens π 192 Explore synthetic data benchmark results as an interactive bookshelf
Running 220 FineVision: Open Data is All You Need π 220 A new open-source dataset for training VLMs
Running 90 Scaling FineWeb to 1000+ languages: Step 1: finding signal in 100s of evaluation tasks π 90 Evaluate multilingual models using FineTasks
Running Featured 1.31k FineWeb: decanting the web for the finest text data at scale π· 1.31k Read a detailed overview of the FineWeb webβscale text dataset
Running on CPU Upgrade Featured 3.05k The Smol Training Playbook π 3.05k The secrets to building world-class LLMs
meituan-longcat/LongCat-Flash-Chat Text Generation β’ 562B β’ Updated Sep 24, 2025 β’ 36.4k β’ 527
Running 79 Maintain the unmaintainable π 79 Explore the complex relationships between 400+ machine learning models
Running on Zero Featured 269 granite-docling-258M demo π 269 Extract and convert document content from images