ConardLi/easy-dataset 1.2.1

[1.2.1] 2025-03-23

on GitHub

🔧 修复

1. 修复文本块排序不准确的问题

⚡ 优化

1. 下调默认并发量为 3 （解决触发部分模型限流问题）
2. 优化问题生成提示词，提升问题生成质量
3. 下调最小分割字符数为 100，上调最大分割字符数为 10000
4. 当模型未按标准格式输出时，日志增加原始输出信息

✨ 新功能

1. 支持编辑问题、自定义问题
2. 支持数据集直接在 LLaMa Factory 中使用
3. 支持配置用户自定义提示词

🔧 Fixes

1. Fixed the inaccurate text block sorting issue.

⚡ Optimizations

1. Reduced the default concurrency to 3 (to solve the problem of triggering rate limits for some models).
2. Optimized the problem generation prompt to improve the quality of generated problems.
3. Reduced the minimum split character count to 100 and increased the maximum split character count to 10,000.
4. Added the original output information to the log when the model does not output in the standard format.

✨ New Features

1. Supported problem editing and custom problem creation.
2. Supported direct use of datasets in LLaMa Factory.
3. Supported configuration of user - defined prompts.