Hello Pickaxe team !
Is it possible to implement the latest models of Mistral ? Medium 3.1 & Magistral medium 1.2
From what I understand Mistral focuses its efforts on tiny/small/medium parameters LLM because of a lack of chips in their possession to train a good large parameters. (They are currently building a datacenter for large 3.)
If you have the time for my accessory request it would be super nice !
(And congrats for your fundraising !!)