config.json

{
  "architectures": [
    "MiniLLaDA"
  ],
  "attention_probs_dropout_prob": 0.1,
  "auto_map": {
    "AutoConfig": "configuration_mini_llada.MiniLLaDAConfig",
    "AutoModel": "modeling_mini_llada.MiniLLaDA"
  },
  "backbone_config": {
    "_name_or_path": "klue/roberta-large",
    "architectures": [
      "RobertaForMaskedLM"
    ],
    "attention_probs_dropout_prob": 0.1,
    "classifier_dropout": null,
    "gradient_checkpointing": false,
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.1,
    "hidden_size": 1024,
    "initializer_range": 0.02,
    "intermediate_size": 4096,
    "layer_norm_eps": 1e-05,
    "max_position_embeddings": 514,
    "model_type": "roberta",
    "num_attention_heads": 16,
    "num_hidden_layers": 24,
    "position_embedding_type": "absolute",
    "tokenizer_class": "BertTokenizer",
    "type_vocab_size": 1,
    "use_cache": true,
    "vocab_size": 32000
  },
  "backbone_model_name": "klue/roberta-large",
  "bos_token_id": 0,
  "classifier_dropout": null,
  "dtype": "float32",
  "eos_token_id": 2,
  "gradient_checkpointing": false,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 1024,
  "initializer_range": 0.02,
  "intermediate_size": 4096,
  "layer_norm_eps": 1e-05,
  "mask_token_id": 4,
  "max_position_embeddings": 514,
  "model_type": "mini-llada",
  "num_attention_heads": 16,
  "num_hidden_layers": 24,
  "pad_token_id": 1,
  "position_embedding_type": "absolute",
  "tokenizer_class": "BertTokenizer",
  "transformers_version": "4.57.3",
  "type_vocab_size": 1,
  "use_cache": true,
  "vocab_size": 32000
}