Update README.md

README.md

@@ -1,7 +1,8 @@
 This model used hfl/chinese-roberta-wwm-ext-large backbone and was trained on SNLI, MNLI, DNLI, KvPI data in Chinese version.
 Model structures are as follows:
 
-`class RobertaForSequenceClassification(nn.Module):
+```
+class RobertaForSequenceClassification(nn.Module):
     def __init__(self, tagset_size):
         super(RobertaForSequenceClassification, self).__init__()
         self.tagset_size = tagset_size
@@ -34,4 +35,4 @@ class RobertaClassificationHead(nn.Module):
         x = self.out_proj(x)
         return x
 model = RobertaForSequenceClassification(num_labels)
-model.load_state_dict(torch.load(args.model_save_path+'Roberta_large_model.pt', map_location=device))`
+model.load_state_dict(torch.load(args.model_save_path+'Roberta_large_model.pt', map_location=device))```