XLM-RoBERTa masked-LM architecture
The captured configuration identifies XLMRobertaForMaskedLM with model type xlm-roberta and Transformers support.
Open Source Model Profile · bonadossou
afrolm_active_learning
afrolm_active_learning is an XLM-RoBERTa-family fill-mask model from bonadossou. According to the model card, it is AfroLM pretrained on 23 African languages.
- Publisher
- bonadossou
- Task
- fill-mask
- Model type
- xlm-roberta
- License
- Unknown
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
afrolm_active_learning is published by bonadossou as a fill-mask model. The captured configuration identifies XLMRobertaForMaskedLM with model type xlm-roberta. According to the model card, it is the AfroLM model pretrained from scratch on 23 African languages for the associated EMNLP 2022 paper.
Recorded capabilities
23-African-language coverage
According to the model card, AfroLM was pretrained from scratch on 23 African languages, with the publisher listing Amharic, Oromo, Bambara, Hausa, Igbo, Swahili, Yoruba, Zulu, and others.
Publisher-reported NER and classification results
According to the model card, the publisher reports MasakhaNER, text-classification, and sentiment-analysis figures alongside AfriBERTa, mBERT, XLMR-base, and AfroXLMR numbers.
Documented XLMRobertaTokenizer usage
According to the model card, the publisher recommends XLMRobertaTokenizer directly, notes a 256 model_max_length setting, and documents Transformers loading code.
Use cases in the source record
- Masked-language-modeling research on the publisher-listed African languages.
- Named-entity recognition, text-classification, and sentiment-analysis experiments consistent with the publisher-reported evaluation datasets.
Limitations and unknowns
- No parameter count was extracted from this record.
- No evaluation results were extracted in structured form; the card table is a publisher claim and was not independently verified.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
Source and provenance
Source: bonadossou/afrolm_active_learning
Captured: Unknown. Processed: 2026-09-07T19:34:41.877151+00:00.
AfroLM: A Self-Active Learning-based Multilingual Pretrained Language Model for 23 African Languages GitHub Repository of the Paper This repository contains the model for our paper AfroLM: A Self-Active Learning-based Multilingual Pretrained Language Model for 23 African Languages which will appear at the Third Simple and Efficient Natural Language Processing, at EMNLP 2022. Our self-active learning framework Languages Covered AfroLM has been pretrained from scratch on 23 African Languages: Amharic, Afan Oromo, Bambara, Ghomalá, Éwé, Fon, Hausa, Ìgbò, Kinyarwanda, Lingala, Luganda, Luo, Mooré, Chewa, Naija, Shona, Swahili, Setswana,…
F001F002F003F004F005F006F007F009F010F011F012F013F014