It is both cost efficient in training (+ future fine-tuning) as well as inferenc... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		tosh on April 24, 2024 \| parent \| context \| favorite \| on: Snowflake Arctic Instruct (128x3B MoE), largest op... It is both cost efficient in training (+ future fine-tuning) as well as inference compared to most other current models. Can you elaborate?

rajhans on April 24, 2024 | [–]

We have published some insights here. https://medium.com/snowflake/snowflake-arctic-cookbook-serie...

ru552 on April 24, 2024 | [–]

the unquantized model is almost 1tb in size and the benchmarks provided by Snowflake shows performance in the middle of the pack compared to other recent releases.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact