ontocord/codepep
收藏Hugging Face2024-10-09 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/ontocord/codepep
下载链接
链接失效反馈官方服务：
资源简介：
---
license: apache-2.0
---
Based on tomekkorbak/codeparrot-clean-train-v2-pep8 and codeparrot/codeparrot-clean. We create a combined `quality_score` as described below: All open source non-GPL code.
Score is computed as follows:
- min(1.0, int(dat['copies'])/20 + dat['ratio']/10 + dat['alpha_frac']* 0.1 + 0.5*(1-dat['avg_score'])) if avg_score exists
- min(1.0, int(dat['copies'])/20 + dat['ratio']/10 + dat['alpha_frac']* 0.1) otherwise
- avg_score is the pep8 score, lower meaning better.
提供机构：
ontocord
