Bringing Ultra Large AI Text Generation Model to the Small Cellphones

AI is making small cellphones and other machines quite competent and efficient

Large-scale artificial intelligence models have become quite popular over the past couple of years. These models can generate articles that are quite similar to the ones written by humans and that can help software developers complete codes. AI models can generate a series of pictures from a sequence of text descriptions to tell stories. The emergence of large-scale AI models has presented researchers with new opportunities for such undertakings.

The remarkable performance of AI models over a wide range of applications often brings a big budget for hardware and running costs. But fortunately, due to the support from the open-source community, serving large AI models has become quite affordable, efficient, and accessible.

The popular AI model OPT-175B was initially introduced by Meta. Basically, the Open Pretrained Transformer is a language model with 175 billion parameters trained on previously available data sets, to enable more community engagement in understanding new technologies. The social media giant released this new AI model under a non-commercial license clause so that it can be used to maintain its integrity and prevent misuse. The company will only allow access to this model to academic researchers, or those affiliated with organizations in government, civil society, and academia, along with industry research laboratories around the world.

The introduction of these large AI language models is about to make small cell phones sturdier and smarter. Besides OPT-175B, other models like GPT-3 and AlphaFold exceed the capacity of existing hardware and complex and distributed technologies that can be used for training and deploying inference services. Furthermore, the emergence of Colossal-AI is also helping machines to efficiently and rapidly streamline artificial intelligence and promote low-cost application and implementation of big models.