Running 13 Defeating the trainer-generator precision mismatch in TRL 🎯 13 Download research PDF (Pro access required)
Running Featured 72 Distilling 100B+ Models 40x Faster with TRL 📝 72 TRL distillation for 100B+ teachers, 40x faster
nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 Text Generation • 124B • Updated 13 days ago • 611k • 338
Running on CPU Upgrade 227 The Synthetic Data Playbook: Generating Trillions of the Finest Tokens 📝 227 Explore synthetic data experiments on a virtual bookshelf