jhu-clsp/mmBERT-decay-data download history
jhu-clsp/mmBERT-decay-data is a fill mask dataset on the Hugging Face Hub. In the last 30 days it was downloaded 18,275 times (4,698 in the last 7 days), and 207,134 times in total. It ranks #1,703 among datasets by monthly downloads.
MMBERT Decay Phase Data Phase 3 of 3: Annealed language learning decay phase (100B tokens) with massive multilingual expansion to 1833 languages. 📊 Data Composition NOTE: there are multiple decay data mixtures: this mixture described below is the Decay-Cont mixture. Howe
Spaces using mmBERT-decay-data
- Model Pulse 8 likes