Datasets:

Modalities:
Text
Formats:
json
Libraries:
Datasets
pandas
License:
huu-ontocord commited on
Commit
2f83b36
1 Parent(s): ead41d5

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -2,7 +2,7 @@
2
  license: apache-2.0
3
  ---
4
 
5
- Based on tomekkorbak/codeparrot-clean-train-v2-pep8. We create a combined score based on codeparrot/codeparrot-clean. All open source non-GPL code.
6
  Score is computed as follows:
7
  - 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
8
  - min(1.0, int(dat['copies'])/20 + dat['ratio']/10 + dat['alpha_frac']* 0.1) otherwise
 
2
  license: apache-2.0
3
  ---
4
 
5
+ 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.
6
  Score is computed as follows:
7
  - 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
8
  - min(1.0, int(dat['copies'])/20 + dat['ratio']/10 + dat['alpha_frac']* 0.1) otherwise