AI Startup Cerebras Releases New GPT Models Similar to ChatGPT
STARTUP TECHNOLOGY 

AI Startup Cerebras Releases New GPT Models Similar to ChatGPT

Administrator

On Tuesday, Cerebras releases new GPT models as like ChatGPT, an open-source models which have been made available to the research and business communities by an artificial intelligence startup.

 Seven models, from tiny 111 million parameter language models to a bigger 13 billion parameter model, were all made available by Silicon Valley-based Cerebras. They were all trained on its Andromeda AI supercomputer.

According to Andrew Feldman, founder and CEO of Cerebras, “there is a major drive to close what has been open sourced in AI…not it’s unexpected as there’s now huge money in it”. “Everything has been so open, and that has contributed greatly to the excitement in the community and the success we’ve made”.

Models with additional parameters can carry out more intricate generating operations.

For instance, ChatGPT an OpenAI chatbot that was released late last year has 175 billion parameters can produce research and poetry and has contributed to generate significant interest in and financing for AI more generally.

While larger models run on PCs or servers, Cerebras claimed that smaller models can be used on smartphones or smart devices, however complicated activities like summarising lengthy passages call for larger models.

If you give a smaller model more training, it can become accurate, according to some interesting publications that have been published, said by Freund. So, there is a trade-off between size and skill.

Due to the design of the Cerebras system, which incorporates a chip the size of a dinner plate designed for AI training, Feldman said that his largest model only required a little more than a week to train. This is a significant improvement over the several months that it generally takes.

Today’s AI models are primarily developed on Nvidia Corp. chips, but more and more companies, including Cerebras, are vying for market dominance.

According to Feldman, the models developed on Cerebras machines can also be customised or further trained on Nvidia systems.