{
    "architecture": "Transformer Encoder-Decoder",
    "src_vocab_size": 50259,
    "tgt_vocab_size": 50259,
    "src_seq_len": 64,
    "tgt_seq_len": 96,
    "d_model": 64,
    "N": 2,
    "h": 4,
    "dropout": 0,
    "d_ff": 128,
    "tokenizer": "gpt2",
    "best_epoch": 6,
    "validation_loss": 2.017836697101593
}