TextSynth access to some trained models that are a neural network of up to 1.5 billion parameters, for example, GPT-J, GPT-NeoX, and M2M100.
We can use this technology for Intelligent customer service, converting text to speech, generating related random text according to your input, asking and answering questions, playing games, etc. TextSynth can custom inference framework it will take advantage of CPU and GPU, then, it will reduce the cost of this service.
By the way, ttsmp3’s(https://ttsmp3.net) service is based on this AI technology. Very simple to use. And it’s totally free for everyone, enjoy~