AI in Neurai

Neurai will leverage various AI models for the applications that will be incorporated into the network. IoT devices will have access to them through an internal medium, executing the AI algorithm directly from their own devices running AI locally, or even accessing AI in the cloud, taking advantage of the power these can offer.

Each type of model must serve a purpose, and each IoT device configuration will have needs that one of these types will meet.

Here we present some of the models with which we will be working throughout the development

AI in IoT equipamment

Artificial Intelligence has evolved from large-scale systems to smaller devices, such as microcontrollers, thanks to specific software techniques and optimizations. This advancement allows low-power devices with limited resources to perform sophisticated data processing tasks.

TensorFlow lite

Developed by Google Brain, TensorFlow is one of the most popular open-source libraries for machine learning and artificial intelligence. It provides a comprehensive ecosystem of tools, libraries, and resources that allow developers to build and deploy machine learning models.

TinyML

This is a subfield of machine learning and an open-source library, specifically designed for microcontrollers like those used in IoT devices. It’s closely related to TensorFlow and enables machine learning on tiny, low-power devices.

ONNX

This is an open-source artificial intelligence ecosystem that allows developers to move models between different AI frameworks, such as TensorFlow, PyTorch, MXNet, and others.

Pytorch

PyTorch is an open-source machine learning library for Python, developed primarily by Facebook’s artificial intelligence research group. It’s known for its flexibility and efficiency, and is widely used in the AI research community.

AI in own computer

Artificial Intelligence has evolved from large-scale systems to smaller devices, such as microcontrollers, thanks to specific software techniques and optimizations. This advancement allows low-power devices with limited resources to perform sophisticated data processing tasks.

Codellama

Codellama is based around a learning system called spaced repetition (or distributed practice), in which problems are revisited at an increasing interval as you continue to progress. In other words, the more you get a problem right, the less frequently you’ll review it and vice-versa

Bert

Google BERT is an algorithm that increases the search engine’s understanding of human language.

This is essential in the universe of searches since people express themselves spontaneously in search terms and page contents.

GPT-2

Generative Pre-trained Transformer 2 (GPT-2) is a large language model by OpenAI and the second in their foundational series of GPT models.

Distilbert

DistilBERT is a small, fast, cheap and light Transformer model trained by distilling BERT base.

AI with API connection

Artificial Intelligence has evolved from large-scale systems to smaller devices, such as microcontrollers, thanks to specific software techniques and optimizations. This advancement allows low-power devices with limited resources to perform sophisticated data processing tasks.

Openai

OpenAI researches artificial intelligence with the declared intention of developing "safe and beneficial" artificial general intelligence, which it defines as "highly autonomous systems that outperform humans at most economically valuable work"

Bard

Bard is a conversational generative artificial intelligence chatbot developed by Google, based initially on the LaMDA family of large language models (LLMs) and later PaLM.

Amazon AI

AI Services easily integrate with your applications to address common use cases such as personalized recommendations, modernizing your contact center, improving safety and security, and increasing customer engagement.

Midjourney

An independent research lab exploring new mediums of thought and expanding the imaginative powers of the human species.

Ollama - LLM Example in local

Linux

curl https://ollama.ai/install.sh | sh

These are some free models that can be used. For starters, Mistral works very well and it is possible to ask questions in several languages.

ollama run mistral
ollama run llama2
ollama run codellama
ollama run llama2-uncensored
ollama run orca-mini
ollama run wizardcoder
ollama run zephyr
ollama run llama2-chinese
ollama run falcon

more soon...