Currently, there is no GPT2-Large or GPT2-XL model that was trained from scratch for Portuguese on the hub: Hugging Face – The AI community building the future. . For this project, the goal is to create a strong language generation model for Portuguese.
A randomly initialized GPT2-Large (and/or GPT2-XL, if we have resources for this ) model
A causal language modeling script for Flax is available here. It can be used pretty much without any required code changes.
The desired project output is a GPT2 model that is able to generate Portuguese language. A nice generation demo can be created for this.
The most important read would be the following colab: