About NeuraNET

NeuraNET started when it's founder, Ruben Roy, wanted to create his own neural network. After spending a few days experimenting with neural networks, he finally created the first version of NeuraNET.

Originally, NeuraNET was a small machine learning model known as Krypt. The model only had 126 thousand parameters, and it was trained on a relatively small corpus.

Ruben then kept expanding on NeuraNET and eventually launched the first real version of the NeuraNET LLM. It was known as NeuraNET Zero. The NeuraNET Zero model was then continuously trained for months and months until it large enough for public release. NeuraNET Hyper was created. It was meant to be a censored, closed-source, and public model that anyone can freely use.

In February 15 2024, NeuraNET was launched. NeuraNET is still growing, and gaining more and more capabilities. It is no longer a language model, but a multimodal AI now along with vision capabilities.