First version of the chinese_roberta_L-12_H-128 model and tokenizer.

README.md

@@ -115,7 +115,7 @@ python3 preprocess.py --corpus_path corpora/cluecorpus.txt \
 ```
 ```
 python3 pretrain.py --dataset_path cluecorpus_seq512_dataset.pt \
-                    --pretrained_model_path 		    models/cluecorpus_roberta_l12h128_seq512_model.bin-1000000 \
+                    --pretrained_model_path models/cluecorpus_roberta_l12h128_seq512_model.bin-1000000 \
 					--vocab_path models/google_zh_vocab.txt \
 					--config_path models/bert_l12h128_config.json \
 					--output_model_path models/cluecorpus_roberta_l12h128_seq512_model.bin \