Δεδομενα εκπαιδευσης
Καλησπερα,
Ευχαριστω που δημοσιευεται το μοντελο, θα ηταν ευκολο να γραψετε περισσοτερα για τα δεδομενα εκπαιδευσης?
Hi,
This is POC project so I doubt you'd find much interest in using it, hence why I left the card blank.
Anyhow, if you are landing here because you have already tried https://huggingface.co/pranaydeeps/Ancient-Greek-BERT, then you are not going to find much better results.
This model is fine tuned on the downstream fill-mask task and was trained on the architecture of the model I linked above.
It's a very small model trained on 31k examples for a total of circa 7k training steps. The training data differs from the model mentioned above in the fact that it does not contain any Byzantine or Hellenistic Greek
Please let me know if you find this useful and if this model is in any way good for what you need. I am working on a prover fine-tuned version of purely classical ancient Greek