Get the latest tech news

Building a Korean ambiguity solver fast enough to skip the GPU: 7,300 words/SEC


How Kimchi Reader's Korean ambiguity solver, a 14M-parameter KoELECTRA quantized to int8, ended up running server-side on a plain CPU with no GPU, resolving thousands of word-sense ambiguities per second. The four attempts it took to get there.

None

Get the Android app

Or read this on Hacker News

Read more on:

Photo of SEC

SEC

Photo of Korean

Korean

Photo of GPU

GPU

Related news:

News photo

NVIDIA challenger D-Matrix claims AI chip 10 times faster than GPU using 5 times less energy, skips DRAM for SRAM

News photo

OpenAI files SEC paperwork to go public

News photo

Confidential submission of draft S-1 to the SEC