config.json

{
  "architectures": [
    "DebertaForMaskedLM"
  ],
  "attention_probs_dropout_prob": 0.1,
  "base_model": "deberta",
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 312,
  "initializer_range": 0.02,
  "intermediate_size": 1200,
  "layer_norm_eps": 1e-07,
  "max_position_embeddings": 512,
  "max_relative_positions": -1,
  "model_type": "deberta",
  "num_attention_heads": 12,
  "num_hidden_layers": 4,
  "pad_token_id": 0,
  "pooler_dropout": 0,
  "pooler_hidden_act": "gelu",
  "pooler_hidden_size": 312,
  "pos_att_type": [
    "c2p",
    "p2c"
  ],
  "position_biased_input": true,
  "relative_attention": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.41.2",
  "type_vocab_size": 0,
  "vocab_size": 14200
}