bert-base-ja / config.json
noriyukipy's picture
Update architecture to BertForPreTraining
565d251
raw
history blame
781 Bytes
{
"_name_or_path": "release/bert-base-ja",
"architectures": [
"BertForPreTraining"
],
"attention_probs_dropout_prob": 0.1,
"bos_token_id": 2,
"cls_token_id": 2,
"eos_token_id": 3,
"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-12,
"mask_token_id": 4,
"max_position_embeddings": 512,
"model_type": "bert",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"sep_token_id": 3,
"tokenizer_class": "DebertaV2Tokenizer",
"transformers_version": "4.8.2",
"type_vocab_size": 2,
"unk_token_id": 1,
"use_cache": true,
"vocab_size": 32000
}