Update README.md

README.md

@@ -1 +1,41 @@
-# ConvLab
+# ConvLab
+This is a toolkit for developing task-oriented dialog system. We
+followed the conventional pipeline framework, where there are 4 seperate
+components: NLU, DST, Policy and NLG.
+
+We offer the base class and some SOTA baseline models (coming soon)
+for each component. Specially, the NLU, DST and NLG models are trained
+individually, while the Policy is trained within a complete pipeline
+system in an RL-based manner.
+
+- NLU: Regex method, Seq2seq, JointNLU, ContextualNLU
+- DST: Rule, NBT, Sequicity
+- Policy: Rule, DQN, MDQN, HRL
+- NLG: Templated method, SC-LSTM, CA-LSTM
+
+## Environment
+
+<!---
+- Trained NLG model can be downloaded [here](https://www.dropbox.com/s/7d6rr57hmdcz9pd/lstm_tanh_%5B1549590993.11%5D_24_28_1000_0.447.pkl?dl=0).
+-->
+- Trained NLG model can be downloaded [here](https://www.dropbox.com/s/u1n8jlgr89jnn2f/lstm_tanh_%5B1552674040.43%5D_7_7_400_0.436.pkl?dl=0). 
+- Trained NLU model can be downloaded [here](https://www.dropbox.com/s/y2aclsz9t7nmxnr/bi_lstm_%5B1552541377.53%5D_7_7_360_0.912.pkl?dl=0).
+- Trained S2S UserSim model can be downloaded [here](https://www.dropbox.com/s/2jxkqp2ad07asps/lstm_%5B1550147645.59%5D_20_29_0.448.p?dl=0).
+- Trained MLST NLU model can be downloaded [here](https://1drv.ms/u/s!AmXaP0QRGLFchVZqB047pJdS-tiT). Unzip the downloaded file in the tasktk/nlu/mlst directory. 
+- Trained JointNLU model can be downloaded [here](https://1drv.ms/u/s!AmXaP0QRGLFchVn7DNj4s7fghLTo). Unzip the downloaded file in the tasktk/nlu/joint_nlu directory. 
+
+## Document
+
+## To Developer
+
+## How to start
+To run the code, you have to first download [mdbt.tar.gz](https://drive.google.com/file/d/1jN8p_PrhgdfBYa2--GqSQiHGFONWuINe/view?usp=sharing)
+ then extract it and move the mdbt directory under ./data . The mdbt directory
+includes the data and trained model required for building MDBT tracker.
+
+Then, you can just run the ./run.py script to run the dialog on dialog-act level.
+Note that the MDBT model receives natural langauge utterances as input, so we used a trivial
+rule-based NLG to convert the user response DA into natural langauge format (see tasktk/nlg/template_nlg.py).
+
+The outputs of system policy, MDBT and simulator are logged into ./session.txt, where the turns and sessions
+are seperated by separators for clarity.