mgpt_config.py

from transformers import PretrainedConfig
from typing import List


class MGPTConfig(PretrainedConfig):
    model_type = "mgpt"

    def __init__(
        self,
        block_size: int = 1024,
        vocab_size: int = 12000,
        n_layer: int = 12,
        n_head: int = 8,
        n_embd: int = 512,
        dropout: float = 0.1,
        bias: bool = False,
        **kwargs,
    ):
        self.block_size = block_size
        self.vocab_size = vocab_size 
        self.n_layer = n_layer
        self.n_head = n_head
        self.n_embd = n_embd
        self.dropout = dropout
        self.bias = bias 
        super().__init__(**kwargs)