config.json

{
  "_name_or_path": "mitre_913m",
  "activation_dropout": 0.0,
  "activation_function": "relu",
  "architectures": [
    "MitreForConditionalGeneration"
  ],
  "attention_dropout": 0.1,
  "auto_map": {
    "AutoConfig": "configuration_mitre.MitreConfig",
    "AutoModel": "modeling_mitre.MitreForConditionalGeneration"
  },
  "bos_token_id": 0,
  "d_model": 1280,
  "decoder_attention_heads": 20,
  "decoder_ffn_dim": 5120,
  "decoder_layers": 36,
  "decoder_start_token_id": 2,
  "dropout": 0.1,
  "eos_token_id": 2,
  "init_std": 0.02,
  "max_position_embeddings": 256,
  "model_type": "mitre",
  "num_hidden_layers": 36,
  "pad_token_id": 1,
  "scale_embedding": true,
  "torch_dtype": "float32",
  "transformers_version": "4.47.1",
  "use_cache": true,
  "vocab_size": 160025
}