README.md

metadata

language:
  - uz
tags:
  - transformers
  - mit
  - robert
  - uzrobert
  - uzbek
  - cyrillic
  - latin
license: apache-2.0
widget:
  - text: Kuchli yomg‘irlar tufayli bir qator <mask> kuchli sel oqishi kuzatildi.
    example_title: Latin script
  - text: >-
      Алишер Навоий – улуғ ўзбек ва бошқа туркий халқларнинг <mask>, мутафаккири
      ва давлат арбоби бўлган.
    example_title: Cyrillic script

UzRoBerta model.

Pre-prepared model in Uzbek (Cyrillic and latin script) to model the masked language and predict the next sentences.

How to use. You can use this model directly with a pipeline for masked language modeling: $$fromtransformersimportpipelineunmasker=pipeline{(}^{\mathrm{\prime }}fill-mas{k}^{\mathrm{\prime }},model{=}^{\mathrm{\prime }}coppercitylabs\mathrm{/}uzbert-base-uncase{d}^{\mathrm{\prime }})unmasker(\mathrm{"}\text{АлишерНавоий–улуғўзбеквабошқатуркийхалқларнинг}[MASK],\text{мутафаккиривадавлатарбобибўлган}\mathrm{.}\mathrm{"})from\; transformers\; import\; pipeline\; unmasker\; =\; pipeline(\text{'}fill-mask\text{'},\; model=\text{'}coppercitylabs/uzbert-base-uncased\text{'})\; unmasker("Алишер\; Навоий\; –\; улуғ\; ўзбек\; ва\; бошқа\; туркий\; халқларнинг\; [MASK],\; мутафаккири\; ва\; давлат\; арбоби\; бўлган.")$$

Training data.

UzBERT model was pretrained on ≈2M news articles (≈3Gb).