PalimpsestBERT

Fresh ModernBERT pretrained Ancient Greek model on a new Ancient Greek corpus derived from the Harvard institutional-books collection.

Pretraining

  • Best val loss: 1.8140, PPL: 6.135 (epoch 39/40)

Downstream (full results in the paper/repo)

TAG POS MORPH LEMMA UAS LAS SENT_F
97.63 97.36 92.38 88.64 80.64 75.77 59.37
Downloads last month
62
Safetensors
Model size
0.1B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Dataset used to train pnadel/palimpsest-modernbert