Genesis Ai

There are only 2 possible reasons to delay weights: 1. It sucks 2. You stole something and want to know if ppl can figure it out You can jailbreak any open source model with some simple trained kv-injections so "safety" is bs.

I prefer code, but this one is an exception.

One of the most efficient ways of wasting compute is to use JSON in LLMs. Not to mention the degradation of perplexity... Don't take my word for it tho, just inspect the attention weights when using MCPs or force the output to follow a JSON schema.

@komplexkonjugat Äntligen, dags att kasta ut egna taffliga cuda kernels för detta

NVIDIA: "We just made scikit-learn, UMAP, and HDBSCAN run on GPUs with zero code changes!" Nice! UMAP goes brrrrrr! reddit.com/r/MachineLearn…

Got a really stupid idea this morning BUT seems its possible to solve arithmetic in ML by just tokenizing smarter and do selective activations. It also solves strawberrry out of the box.

@_carlhannes @komplexkonjugat @J_Landstroem Tar med nästa gång så testar vi

@_carlhannes @komplexkonjugat @J_Landstroem Bruh jag har ju risers, iofs 1x-16x men endån.

Perplexity släpper en avcensurerad open-soucre variant av DeepSeek-R1 (och kinesiska troll blir sura). Den går nu att även att nå via ollama. Även en distill är släppt, men på 70b. Det tar nog inte lång tid innan vi har mindre distills. perplexity.ai/hub/blog/open-…

@danielhanchen @UnslothAI That was fun! Got a nf4 fused dequant kernel to x1.31 speedup at least with the given constrains.

We made 5 challenges and if you score 47 points we'll offer you $500K/year + equity to join us at 🦥@UnslothAI! No experience or PhD needed. $400K - $500K/yr: Founding Engineer (47 points) $250K - $300K/yr: ML Engineer (32 points) Challenges: 1. Convert nf4 / BnB 4bit to Triton 2. Make FSDP2 work with QLoRA 3. Remove graph breaks in torch.compile 4. Help solve Unsloth issues! 5. Memory Efficient Backprop If you have any questions about the challenges, please feel free to ask! We're looking for people to help push Unsloth forward - so come join us to democratize AI further! Our past work includes: 1. 1.58bit DeepSeek R1 GGUFs: x.com/UnslothAI/stat… 2. GRPO with Llama 3.1 8B in a Colab: x.com/UnslothAI/stat… 3. Gemma bug fixes: x.com/danielhanchen/… 4. Gradient accumulation bug fixes: x.com/danielhanchen/… Details & submission guide: colab.research.google.com/drive/1JqKqA1X…

@UnslothAI @danielhanchen @UnslothAI was asking for a x1.15 speedup, I give you x1.31 💃 aaaand works with torch.compile, triton autotune, T4 gpus or just like these benchmarks, out of the box. Still have some more tricks on optimizing it but that is for another night! Also should do the MM in there.

@UnslothAI @danielhanchen 3 hours in, need to wrap it up now. Just some last optimizations and then showtime!

I think its time for a hacknight! @UnslothAI makes good kernels so lets try their challenge. Always start with the hard ones right? Lets start with a fused nf4 tensor kernel in Triton!

@_carlhannes @JoakimEwenson @0x4a45 @jhakansson_ Det är ditt samvete som pratar, du vet att du kan göra det där med typ hälften av kod och dubbelt så effektivt. Är det rimligt? Troligen inte, bra jobbat kompis ❤️

@JoakimEwenson @0x4a45 @jhakansson_ Väntar bara på att @_genesis_ai_ Ska codegolfa ut mig här och göra en egen binary på typ 5 kb som har samma funktionalitet

har ni någon gång blivit så arga på kubernetes att ni bygger er egna templatingtooling det har jag npmjs.com/package/picohe…

2 år sedan, börjar ju endå bli dags för en remake med senaste modellerna