yuyou-dev / chatgpt-fine-tuning

Quick-start guide to fine-tuning ChatGPT using Python. Includes scripts for data preprocessing, model training, and evaluation. 快速入门指南: 使用Python微调ChatGPT。包含数据预处理、模型训练和评估脚本。

ChatGPT-Fine-tuning / ChatGPT微调指南

Quick-start guide to fine-tuning ChatGPT using Python. Includes scripts for data preprocessing, model training, and evaluation.

快速入门指南: 使用Python微调ChatGPT。包含数据预处理、模型训练和评估脚本。

Project Structure / 项目结构

chatgpt-finetuning/
│
├── data/
│   ├── raw/
│   │   ├── training_data.txt
│   │   └── validation_data.txt
│   └── processed/
│       └── ...
│
├── scripts/
│   ├── preprocess_data.py
│   ├── upload_data.py
│   └── finetuning.py
│
├── notebooks/
│   ├── fine-tuning.ipynb
│   ├── fine-tuning-format.ipynb
│   ├── fine-tuning-format-comparison.ipynb
│   └── 视频解说.ipynb
│
├── README.md
└── requirements.txt

Getting Started / 开始

1. Clone the Repository / 克隆仓库:

git clone https://github.com/yuyou-dev/chatgpt-fine-tuning.git
cd chatgpt-fine-tuning

2. Install Dependencies / 安装依赖:

pip install -r requirements.txt

3. Fine-tuning the Model / 微调模型:

Navigate to the scripts/ directory and run the script to preprocess data:

Preprocess Data / 预处理数据

python3 preprocess_data.py

Upload Data / 上传数据

Run the upload script:

python3 upload_data.py

Fine-tuning / 微调训练

Run the fine-tuning script:

python3 finetuning.py

Running Tests / 验证和测试

Run API test on PlayGround

Colab 微调

• 1.微调基础用例
• 2.微调实战-训练指定格式和字段的输出
• 3.微调实战-训练指定格式和字段的输出-不同方式的对比

其他GPT技术

• 1.GPT-4v实现AI视频解说 - Colab)

Contributing / 贡献

We welcome contributions! Please open an issue or submit a pull request on GitHub. 我们欢迎各种贡献！请在GitHub上开启一个问题(issue)或提交一个拉取请求(pull request)。

License / 许可

MIT License

chatgpt-fine-tuning's Issues

视频解说

APIStatusError Traceback (most recent call last)
in <cell line: 28>()
26 ]
27
---> 28 result = client.chat.completions.create(
29 model="gpt-4-vision-preview",
30 messages=PROMPT_MESSAGES,

4 frames
/usr/local/lib/python3.10/dist-packages/openai/_base_client.py in _request(self, cast_to, options, remaining_retries, stream, stream_cls)
947
948 log.debug("Re-raising status error")
--> 949 raise self._make_status_error_from_response(err.response) from None
950
951 return self._process_response(

APIStatusError: Error code: 307

当时是使用demo , 一直报这个错误。
或许我用的中转key 问题？