A Large-scale Chinese Short-Text Conversation Dataset and Chinese pre-training dialog models - View it on GitHub