bigscience-workshop/metadata
Experiments on including metadata such as URLs, timestamps, website descriptions and HTML tags during pretraining.
PythonApache-2.0
Watchers
- bsagotInria
- dieuwkehupkesUvA
- eemailme
- epavlick
- ggdupontEurope
- huu4ontocord
- justheuristicYSDA
- lab8-tomofiOsaka, Japan
- manandeySalesforce
- mbofb
- norakassner
- SaulLuHugging Face
- schwabdidierUniv. Grenoble Alpes
- shanyas10Walmart Labs
- stephenbachBrown University
- timoschickMunich, Germany
- VictorSanh@huggingface
- yjerniteCIMS, NYU