Feature/quantization/langchain——can run local get the same result of big model

[HardSoft2023]

1、simplfy the langchai_qa.py demo code so that it can run locally

2、python langchain_qa_quantization.py --model_path path/to/chinese_alpaca_plus_pth/7B/ggml-model-q4_0.bin --file_path doc.txt