Samuel Zeng
@SamuelZengML
Founder, | On-device foundation models. High intelligence, low memory, low power | MIT TR35
Joined October 2012
235 Following 3.7K Followers
Something big is coming on September 10. A major open-source release from us.
35B parameters. One iPhone. No cloud.
We trained Edge8-35B, an ultra-sparse MoE with a jointly trained dynamic expert planner, and built an SSD-streaming inference engine around it.
In this demo: 44 tok/s, ~1.06 GB peak memory.
A truly usable large-model stack for on-device AI.
Model, runtime, and paper: open source soon.
Show more
