readme

Fan, Kai · Fan, Kai · commit 244bf5ba4680 · 2018-12-08T22:23:44.000-08:00
diff --git a/README.md b/README.md
@@ -11,7 +11,7 @@ This repository provides an unofficial released implementation of paper ["Biling
 2. Preprocessing including tokenization, lowercasing, and vocabulary files.
 3. The parallel data can be stored in `data/para`, and the example vocab files are in folder `data/vocab`.
 4. Run `./expert_train.sh` to train bilingual expert model, and due to the large dataset, we provide the multi GPU implementation.
-5. Download the [QE dataset](https://lindat.mff.cuni.cz/repository/xmlui/handle/11372/LRT-2619). An example dataset of De-En has been downloaded and preprocessed in folder `data/qe`.
+5. Download the [QE dataset](https://lindat.mff.cuni.cz/repository/xmlui/handle/11372/LRT-2619). An example dataset of sentence level De-En QE task has been downloaded and preprocessed in folder `data/qe`, including human features (If no human feature is prepared, set the argument `--use_hf=False`). 
 6. Run `./qe_train.sh` to train the quality estimation model, and due to the small dataset, we only provide the single GPU implementation.
 7. Run `./qe_infer.sh` to make the inference on dataset without labels.
 
