import torch
import torch.nn as nn


class Brain(nn.Module):
    """
    Minimal Transformer-based autoregressive model.
    """

    def __init__(
        self,
        vocab_size: int,
        d_model: int = 256,
        nhead: int = 4,
        num_layers: int = 2,
        dim_feedforward: int = 512,
        max_seq_len: int = 128,
    ):
        super().__init__()
        self.token_emb = nn.Embedding(vocab_size, d_model)
        self.pos_emb = nn.Parameter(torch.zeros(1, max_seq_len, d_model))
        encoder_layer = nn.TransformerEncoderLayer(
            d_model=d_model,
            nhead=nhead,
            dim_feedforward=dim_feedforward,
            batch_first=True,
        )
        self.transformer = nn.TransformerEncoder(encoder_layer, num_layers)
        self.fc_out = nn.Linear(d_model, vocab_size)
        self.max_seq_len = max_seq_len

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        seq_len = x.size(1)
        x = self.token_emb(x) + self.pos_emb[:, :seq_len, :]
        x = self.transformer(x)
        return self.fc_out(x)