{
  "architectures": [
    "LUAR"
  ],
  "auto_map": {
    "AutoConfig": "config.LUARConfig",
    "AutoModel": "model.LUAR"
  },
  "embedding_size": 512,
  "k_bucket_size": 1024,
  "upstream_transformer_revision": null,
  "model_type": "LUAR",
  "q_bucket_size": 512,
  "torch_dtype": "float32",
  "transformers_version": "4.33.2",
  "use_memory_efficient_attention": false,
  "backbone_config": {
    "_name_or_path": "old_models/paraphrase-distilroberta-base-v1/0_Transformer",
    "architectures": [
      "RobertaModel"
    ],
    "attention_probs_dropout_prob": 0.1,
    "bos_token_id": 0,
    "eos_token_id": 2,
    "gradient_checkpointing": false,
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.1,
    "hidden_size": 768,
    "initializer_range": 0.02,
    "intermediate_size": 3072,
    "layer_norm_eps": 1e-05,
    "max_position_embeddings": 514,
    "model_type": "roberta",
    "num_attention_heads": 12,
    "num_hidden_layers": 6,
    "output_hidden_states": true,
    "pad_token_id": 1,
    "position_embedding_type": "absolute",
    "transformers_version": "4.7.0",
    "type_vocab_size": 1,
    "use_cache": true,
    "vocab_size": 50265
  }
}