🏗️ Building on HF

David Davidich

Ddavidich

·

DDavidich12

AI & ML interests

Making long context actually work on one consumer GPU · Hybrid linear-attention models

Recent Activity

updated a model about 2 months ago

Ddavidich/LOMONOSOV-ZENIT-27B-1M-INDEV

updated a collection about 2 months ago

updated a Space about 2 months ago

Ddavidich/lomonosov-zenit-altay

View all activity

Organizations

Articles 1

Article

Four things that did not fix long-context retrieval, and the one measurement that explained why

Collections 1

spaces 1

LOMONOSOV ZENIT 27B ALTAY

A million tokens of context on one consumer GPU — measured

models 1

Ddavidich/LOMONOSOV-ZENIT-27B-1M-INDEV

Text Generation • 16B • Updated Jul 30 • 34 • 1

datasets 0

None public yet