AttributeError: 'NoneType' object has no attribute 'config'

[KangChou]

• UBUNTU16.04
• PYTHON3.6

of word in train: 55304:

of n-gram in memory: 71499

Traceback (most recent call last): File "wmseg_main.py", line 677, in main() File "wmseg_main.py", line 667, in main train(args) File "wmseg_main.py", line 101, in train seg_model = WMSeg(word2id, gram2id, label_map, hpara, args) File "/data/my_project/nlp_text/WMSeg/wmseg_model.py", line 98, in init self.hpara['config'] = self.bert.config AttributeError: 'NoneType' object has no attribute 'config'

[YingZhuY]

你需要在 bert 预训练模型同目录下加个 config.json 文件

比如我用的 bert-base-chinese 就是

{
  "attention_probs_dropout_prob": 0.1, 
  "directionality": "bidi", 
  "hidden_act": "gelu", 
  "hidden_dropout_prob": 0.1, 
  "hidden_size": 768, 
  "initializer_range": 0.02, 
  "intermediate_size": 3072, 
  "max_position_embeddings": 512, 
  "num_attention_heads": 12, 
  "num_hidden_layers": 12, 
  "pooler_fc_size": 768, 
  "pooler_num_attention_heads": 12, 
  "pooler_num_fc_layers": 3, 
  "pooler_size_per_head": 128, 
  "pooler_type": "first_token_transform", 
  "type_vocab_size": 2, 
  "vocab_size": 21128
}

具体方法可参考 此博客