Faunix
Small, but efficient AI models :)
QwenSeek-2B - distillation of thinking of DeepSeek V4 into Qwen3.5-2B!
We conducted the distillation of thinking from the DeepSeek-V4 model, to Qwen3.5-2B. Now this is your mini DeepSeek!
Our April Fools' AI!
LiarAI-2B — the ultimate liar of 2026. This model hallucinates in approximately 101% of cases... The symbol of April Fools' Day: Trust no one! Now, it is the symbol of LiarAI...
Our April Fools' AI!
YEN, our 258-parameter model that can do a lot, namely, write Yes and No to your queries!
???
???
We are school students studying AI. We are experimenting with fine-tuning models, and in the future, we even plan to train them from scratch.
There are only two of us, two 13-year-old school students! We work on Kaggle T4 GPUs and have already released 4 models. One of them (QwenSeek-2B GGUF version) has already reached over 10,000 downloads!
Actually, we haven't figured out our main goal yet... But for now, our motto is "Small, but efficient AI models". This means we build compact yet effective AI models that can achieve incredible results despite their small size! We are also proving that school students are actively exploring AI — not just by using it, but by creating it :)