Update README.md

README.md

@@ -96,7 +96,7 @@ output = model(encoded_input)
 
 ## Training data
 
-CLUECorpusSmall is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
+[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
 
 ## Training procedure
 
@@ -142,6 +142,13 @@ python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
 					--tie_weights --embedding word_pos_seg --encoder transformer --mask fully_visible --target mlm
 ```
 
+Finally, we convert the pre-trained model into Huggngface's format:
+```
+python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path pytorch_model.bin 
+                                                        --output_model_path huggingface_model.bin 
+                                                        --layers_num 12 --target mlm
+```
+
 ### BibTeX entry and citation info
 
 ```