Update README.md

README.md

@@ -6,34 +6,58 @@ tags:
 - Chinese
 ---
 
-# Open-Chinese-LLaMA-7B-Patch
+# Open-Chinese-LLaMA
 
-This model is a **Chinese large language model base** generated from the [LLaMA](https://github.com/facebookresearch/llama)-7B model after **secondary pre-training** on Chinese datasets.
+[![](https://img.shields.io/github/license/OpenLMLab/OpenChineseLLaMA?label=Code%20License)]()[![Model License](https://img.shields.io/badge/Model%20License-Apache_2.0-green.svg)]()[![](https://img.shields.io/github/last-commit/OpenLMLab/OpenChineseLLaMA)]()[![](https://img.shields.io/github/issues/OpenLMLab/OpenChineseLLaMA)]()
 
-This model is a **patch** model and must be used in conjunction with the official weights. For the installation of the patch and related tutorials, please refer to [OpenLMLab/OpenChineseLLaMA](https://github.com/OpenLMLab/OpenChineseLLaMA).
+This project is a **Chinese large language model base** generated through **incremental pre-training on Chinese datasets** based on [LLaMA](https://github.com/facebookresearch/llama)-7B.
 
-## Usage
+## Features
 
-Since the official weights for [LLaMA](https://github.com/facebookresearch/llama)-7B have not been open-sourced, the model released this time is of the **patch** type, which needs to be used in combination with the original official weights.
+* This project provides a Chinese pre-trained model obtained through full-tuning, including Huggingface version weights.
+* Compared to the original LLaMA, this model has significantly improved Chinese understanding and generation capabilities, achieving outstanding results in various downstream tasks. See [Evaluation](##Evaluation) for details.
+* This project provides tools for converting Huggingface version weights and Meta version weights.
+* Supports [🤗transformers](https://github.com/huggingface/transformers), and provides command-line tools for easy model testing.
 
-You can install the **patch** using `tools/patch_model.py`, for example:
+## Contents
+* [Model Download](##Model%20Download)
+* [Local Demo](##Local%20Demo)
+* [Evaluation](##Evaluation)
+* [Model Format Conversion](##Model%20Format%20Conversion)
 
-```bash
+## Model Download
+
+| Model Name                    | Weight Type | Download Link                                                     | SHA256                 |
+| --------------------------- | -------- | ------------------------------------------------------------ | ---------------------- |
+| Open-Chinese-LLaMA-7B-Patch | Patch    | [[🤗Huggingface]]() <br> [[Baidu Cloud]](https://pan.baidu.com/s/14E7iZKcH-5SHMDu97k70cg?pwd=gk34)<br>[[Google Driver]](https://drive.google.com/drive/folders/1THvuFzq_wojVfMLYV1qsSE_ddSjG0Ypv?usp=sharing) | [SHA256](./SHA256.txt) |
+
+### Usage Notes
+
+
+
+Meta officially released [LLaMA](https://github.com/facebookresearch/llama) does not open-source weights. To comply with relevant licenses, the model released this time is of the **patch** type, and must be used in conjunction with the official original weights.
+
+We provide a [script](https://github.com/OpenLMLab/OpenChineseLLaMA) for installing the **patch**. After obtaining the official weights through regular channels, you can install the patch as follows:
 
+```bash
 python tools/patch_model.py --base_model <path_or_name_to_original_model>
                             --patch_model openlmlab/open-chinese-llama-7b-patch
                             --base_model_format <hf_or_raw>
-
 ```
 
-The **patch** is installed in place, which means that the installed **patch** is the complete `hf` format weight. You can use `transformers` to load the model.
+Note: The installation method of this patch is inplace installation, that is, the installed patch is the complete Huggingface version of this model weight, and you can use transformers to load the model.
 
-## Quick Experience via Command Line
-
-The **patched** model can be easily loaded by `transformers`. For a quick experience, we provide a console Demo:
+Note: This script depends on [OpenLMLab/collie](https://github.com/OpenLMLab/collie), please install this framework using the following command:
 
 ```bash
+pip install git+https://github.com/OpenLMLab/collie.git
+```
 
+## Local Demo
+
+For quick and easy model testing, we provide a command-line version of the demo. After successfully installing the patch according to [Usage Notes](###Usage%20Notes), you can use the script to start an interactive interface:
+
+```bash
 python cli_demo.py --model openlmlab/open-chinese-llama-7b-patch
                    --devices 0
                    --max_length 1024
@@ -42,5 +66,51 @@ python cli_demo.py --model openlmlab/open-chinese-llama-7b-patch
                    --top_p 0.8
                    --temperature 0.7
                    --penalty 1.02
+```
+
+### Examples
+
+Open-Chinese-LLaMA-7B on the left, original LLaMA on the right:
+
+<div align=center><img src="./cli_demo1.png"></div>
+<center style="font-size:14px;color:#C0C0C0;text-decoration:underline">text generation</center>
+<br>
+<div align=center><img src="./cli_demo2.png"></div>
+<center style="font-size:14px;color:#C0C0C0;text-decoration:underline">code generation</center>
+<br>
+<div align=center><img src="./cli_demo3.png"></div>
+<center style="font-size:14px;color:#C0C0C0;text-decoration:underline">instructions (Note: None have been Instruct-tuning)</center>
+<br>
+
+## Evaluation
+
+Open-Chinese-LLaMA-7B performs far better than the original LLaMA on various tasks in Chinese and English datasets. The evaluation results of this model on some datasets are given below (the following indicators are Accuracy, the bigger the better):
+
+| Dataset   | LLAMA 7B | Open-Chinese-LLaMA-7B |
+| -------- | -------- | ----------- |
+| OCNLI    | 31.5     | 45.5        | 
+| CHID     | 25.87    | 71.47       | 
+| TNEWS    | 8.70     | 26.78       | 
+| CMRC     | 11.89    | 34.48       | 
+| PIQA     | 79.8     | 77.31       |
+| HumanEval | 10.5    | 14.63       |
+| MBPP      | 17.7    | 17.2        |
+| **Average**    | 26.57    | 41.05 |
+
+
+Note: See [Benchmark.md](./benchmark/Benchmark.md) for full results
+
+## Model Format Conversion
+
+The model generated by [`patch_model.py`]((https://github.com/OpenLMLab/OpenChineseLLaMA)) in this project is **hf** format which can be loaded by [🤗transformers](https://github.com/huggingface/transformers). For convenience, we also provide a [conversion tool]((https://github.com/OpenLMLab/OpenChineseLLaMA)) between the official version model (raw) and hf:
+
+```bash
+python convert_model.py --model_path <path_or_name_to_your_hf_or_raw_model>
+                        --source_format hf
+                        --target_format raw
+                        --target_path <path_you_want_to_save_the_converted_model>
+                        --raw_parallel_degree 2
+                        --raw_parallel_devices 0,1
+```
 
-```
+Tip: When converting a model in raw format, you need to specify the tensor parallel size and corresponding device, and it can only be converted on a machine with a corresponding number of graphics cards.