Running Featured 1.38k FineWeb: decanting the web for the finest text data at scale π· 1.38k Explore and download the FineWeb webβscale text dataset
Running 118 The Eiffel Tower Llama π 118 Explore the Eiffel Tower Llama experiment with open-source models
Running 115 Unlocking On-Policy Distillation for Any Model Family π 115 Explore on-policy distillation visualization for any model
Running Featured 90 Distilling 100B+ Models 40x Faster with TRL π 90 TRL distillation for 100B+ teachers, 40x faster
VibeVoice Collection Frontier Text-to-Speech Models https://microsoft.github.io/VibeVoice/ β’ 8 items β’ Updated Mar 2 β’ 248
Running on CPU Upgrade Featured 3.22k The Smol Training Playbook π 3.22k The secrets to building world-class LLMs
Running 3.91k The Ultra-Scale Playbook π 3.91k The ultimate guide to training LLM on large GPU Clusters