config.json

{
  "architectures": [
    "Seq2SeqTransformerModel"
  ],
  "d_model": 256,
  "dropout": 0.1,
  "max_seq_length": 75,
  "num_heads": 8,
  "num_layers": 6,
  "torch_dtype": "float32",
  "transformers_version": "4.47.0",
  "vocab_size_src": 15000,
  "vocab_size_tgt": 15000
}