{
  "actv_fn": "gelu",
  "add_token_type_emb": false,
  "architectures": [
    "BertBlocksForMaskedLM"
  ],
  "attn_dropout_prob": 0.1,
  "attn_out_bias": true,
  "attn_proj_bias": true,
  "classifier_dropout": 0.1,
  "emb_dropout_prob": 0.1,
  "global_attention_every_n_layers": 0,
  "head_type": "mlp",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 768,
  "include_final_norm": true,
  "initializer_cutoff_factor": 3.0,
  "initializer_gain": 1.0,
  "initializer_kind": "trunc_normal",
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "local_attention": [
    -1,
    -1
  ],
  "mask_token_id": 1,
  "max_sequence_length": 512,
  "mlp_in_bias": true,
  "mlp_out_bias": true,
  "mlp_type": "mlp",
  "model_type": "bertblocks",
  "norm_bias": true,
  "norm_eps": 1e-12,
  "norm_fn": "layer",
  "norm_kind": "pre",
  "norm_params": {},
  "num_attention_heads": 12,
  "num_blocks": 12,
  "pad_token_id": 0,
  "pos_emb_kind": "alibi",
  "pos_emb_kwargs": {
    "dim": 64,
    "max_seq_len": 512
  },
  "problem_type": "regression",
  "torch_dtype": "float32",
  "transformers_version": "4.53.2",
  "type_vocab_size": 1,
  "unpadding": false,
  "vocab_size": 28996
}