|
--- |
|
language: |
|
- he |
|
tags: |
|
- language model |
|
datasets: |
|
- responsa |
|
--- |
|
|
|
**AlephBERT-base-finetuned-for-shut** |
|
|
|
**Hebrew Language Model** |
|
|
|
Based on alephbert-base: https://huggingface.co/onlplab/alephbert-base#alephbert |
|
|
|
|
|
**How to use:** |
|
|
|
from transformers import AutoModelForMaskedLM, AutoTokenizer |
|
|
|
checkpoint = 'ysnow9876/alephbert-base-finetuned-for-shut' |
|
|
|
tokenizer = AutoTokenizer.from_pretrained(checkpoint) |
|
|
|
model= AutoModelForMaskedLM.from_pretrained(checkpoint) |
|
|
|
#if not finetuning - disable dropout |
|
|
|
model.eval() |
|
|
|
**Training Data** |
|
|
|
about 26,000 different responsa from different rabbis from the past few hundred years |
|
|
