The key code used in this paper is available here. All code will be made publicly available upon the paper's acceptance. The provided code includes implementations for the training process and the main model architectures employed in BrainChat. Below is a description of the files included in the code:

• main.py: The main file containing the initialization and construction details of the primary models used in BrainChat.
• train.py: The file detailing the training processes.
• models.py: The file containing the primary models used in BrainChat, including:
  • FCoCa: BrainChat models.
• loss.py: Contains the detailed implementation of the loss functions used in BrainChat.