Palette: New Series of Small Model Creative Distills With Only The Best Teachers!
One of the two builds for the future Pallete-2.6B-RP-Reasoning-2609-v1.
Teacher model which was used to generate the dataset: GLM 5.3 Flash.
Testing different unique approaches
The A and B stand for different approaches, not different scale of cooking, both baked properly but with slightly different recipes.
Additional information to anybody wants to financially support me and my activity(I already began saving up for the next project):
Solana | EeQMECL51LHoKnpFybjSafpgjBrHyzUcWuo6Z6NEQLHc
Ethereum | 0xb95fe88e4de47fcd120f66a1a27e1e3ee6b8e157
Boosty(Non-Crypto, most convenient) | https://boosty.to/Indexnusrefather