-
DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models
Paper • 2309.14509 • Published • 22 -
LLM Augmented LLMs: Expanding Capabilities through Composition
Paper • 2401.02412 • Published • 38 -
DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models
Paper • 2401.06066 • Published • 59 -
Tuning Language Models by Proxy
Paper • 2401.08565 • Published • 22
Jason Wolosonovich
wolosonovich
AI & ML interests
None yet
Organizations
-
DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models
Paper • 2309.14509 • Published • 22 -
LLM Augmented LLMs: Expanding Capabilities through Composition
Paper • 2401.02412 • Published • 38 -
DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models
Paper • 2401.06066 • Published • 59 -
Tuning Language Models by Proxy
Paper • 2401.08565 • Published • 22
models 0
None public yet
datasets 0
None public yet