3 ms·
It is both cost efficient in training (+ future fine-tuning) as well as inference compared to most other current models. Can you elaborate?
by tosh 3y ago
It is both cost efficient in training (+ future fine-tuning) as well as inference compared to most other current models.
Can you elaborate?
- ru552 3y agothe 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.
- rajhans 3y agoWe have published some insights here. https://medium.com/snowflake/snowflake-arctic-cookbook-series-exploring-mixture-of-experts-moe-c7d6b8f14d16 https://medium.com/snowflake/snowflake-arctic-cookbook-serie...