SpecBundle Collection A collection of production-grade draft models for speculative decoding • 18 items • Updated 3 days ago • 17
Running Agents 1.5k Big Code Models Leaderboard 📈 1.5k Explore and submit code model evaluations on a leaderboard
Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads Paper • 2401.10774 • Published Jan 19, 2024 • 60
Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Models Paper • 2312.06585 • Published Dec 11, 2023 • 29