Running Featured 61 Distilling 100B+ Models 40x Faster with TRL π 61 TRL distillation for 100B+ teachers, 40x faster
Running on CPU Upgrade Featured 3.11k The Smol Training Playbook π 3.11k The secrets to building world-class LLMs
Running 3.79k The Ultra-Scale Playbook π 3.79k The ultimate guide to training LLM on large GPU Clusters