config.json

{
  "_name_or_path": "single_bp_2k_step19999",
  "architectures": [
    "BertForMaskedLM"
  ],
  "attention_probs_dropout_prob": 0.1,
  "attn_norm_layer_type": "layer_norm",
  "attn_num_groups": 1,
  "auto_map": {
    "AutoConfig": "configuring_nt_bert.BertConfig",
    "AutoModelForMaskedLM": "modeling_nt_bert.BertForMaskedLM"
  },
  "classifier_dropout": "None",
  "embedding_norm_layer_type": "layer_norm",
  "embedding_num_groups": 1,
  "embedding_size": 1280,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 1280,
  "initializer_range": 0.02,
  "intermediate_size": 5120,
  "layer_norm_eps": 1e-12,
  "max_position_embeddings": 2000,
  "model_type": "bert",
  "mup": true,
  "num_attention_heads": 16,
  "num_hidden_layers": 24,
  "output_mult": 1,
  "pad_token_id": 3,
  "position_embedding_type": "alibi",
  "prenorm": false,
  "query_zero_init": false,
  "readout_zero_init": false,
  "summary_activation": "gelu",
  "summary_last_dropout": 0.1,
  "summary_type": "first",
  "summary_use_proj": true,
  "torch_dtype": "float32",
  "transformers_version": "4.25.1",
  "type_vocab_size": 2,
  "vocab_size": 10
}