For now we've only tested on PaddlePaddle v1.0, to install PaddlePaddle and for more details about PaddlePaddle, see [PaddlePaddle Homepage](http://paddlepaddle.org).
#### Training
The DuReader model can be trained by run `train.py`, for complete usage run `python train.py -h`.
The DuReader model can be trained by run `run.py`, for complete usage run `python run.py -h`.
The basic training and infering process has been wrapped in `run.sh`, the basic usage is:
```
bash run.sh TASK_NAME
bash run.sh --TASK_NAME
```
For example, to train the model, run:
```
bash run.sh train
bash run.sh --train
```
## Run DuReader on multilingual datasets
To help evaluate the system performance on multilingual datasets, we provide scripts to convert MS MARCO V2 data from its format to DuReader format.
[MS MARCO](http://www.msmarco.org/dataset.aspx)(Microsoft Machine Reading Comprehension) is an English dataset focused on machine reading comprehension and question answering. The design of MS MARCO and DuReader is similiar. It is worthwhile examining the MRC systems on both Chinese (DuReader) and English (MS MARCO) datasets.
You can download MS MARCO V2 data, and run the following scripts to convert the data from MS MARCO V2 format to DuReader format. Then, you can run and evaluate our DuReader baselines or your DuReader systems on MS MARCO data.
#### Inference
To infer a trained model, run the same command as training and change `train` to `infer`, and add `--testset <path_to_testset>` argument. for example, suppose a model is successfully trained and parameters of the model are saved in a directory such as `models/1`, to infer the saved model, run:
