Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
Tianle Cai, Yuhong Li, Zhengyang Geng, Hongwu Peng, Jason D. Lee, Deming Chen, Tri Dao
De quem é
- Carnegie Mellon Universityafiliação declarada pelos autores, segundo o OpenAlex
- University of Illinois Urbana-Champaignafiliação declarada pelos autores, segundo o OpenAlex
Outros papers das mesmas organizações
- Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement LearningGoogle · 1.789 citações
- Medical SAM adapter: Adapting segment anything model for medical image segmentationAgency for Science, Technology and Research · 1.361 citações
- Dynamic 3D Gaussians: Tracking by Persistent Dynamic View SynthesisCarnegie Mellon University · 1.121 citações
- Power Hungry Processing: Watts Driving the Cost of AI Deployment?Carnegie Mellon University · 929 citações
- SplaTAM: Splat, Track & Map 3D Gaussians for Dense RGB-D SLAMCarnegie Mellon University · 921 citações
- VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web TasksCarnegie Mellon University · 912 citações