clue/roberta_chinese_3L312_clue_tiny

With the help ofHuggingface-Transformers 2.5.1, you could use these model as follows: . You could use BertTokenizer.from_pretrained("MODEL_NAME") to train the model . The model was trained on TPU and the details are as follows . It is based on the model of Super_small_sentence-sentence_pair_model and the size of the sentence . The size of this model is 7.5M and the value of each sentence is $7.5 M and the number of sentences is $1.2 M and $2.2M . It was trained using the model on the TPU .,