VEXMLM — TIGQA

Tigrinya question answering fine-tuned from Hailay/VEXMLM, the vocabulary-extended XLM-R for Ge'ez-script languages.

Official implementation: https://github.com/hailaykidu/VEXMLM

This checkpoint performs poorly in absolute terms. Exact Match of 2.39 is near zero. TIGQA's development split contains only 67 questions, so the metric is both very hard and very high-variance. It is published for completeness and reproducibility of the paper's evaluation, not as a usable Tigrinya QA system. For Tigrinya extractive QA, see Hailay/VEXMLM-TiQuAD.

Five-seed benchmark evaluation

Fine-tuned independently under seeds 42–46 with one configuration (hash ce27cc194946) on an A100-PCIE-40GB. Reported as mean ± standard deviation over the five runs, on the dataset's development split (the fine-tuning script evaluates QA on the validation split when one exists).

These are the paper's verified results. They come from the five-seed evaluation described above — not from interactive use.

Interactive inference vs. benchmark

Benchmark evaluation is the five-seed measurement on the held-out development split, shown in the table above.

Interactive inference is what the usage example below performs: Supply a Tigrinya context and question; the model returns an extracted span. Predictions on arbitrary user input are demonstrations only and do not produce or reproduce the benchmark score.

Repository layout

Five independently fine-tuned checkpoints, one per seed. The reported benchmark score is the mean ± standard deviation over all five; no single seed is the "five-seed model."

seed-42/  seed-43/  seed-44/  seed-45/  seed-46/

Load a specific seed with the subfolder argument, as in the example below.

Fine-tuning

Fine-tuned from Hailay/VEXMLM, a vocabulary-extended XLM-R (280,002 subwords, 30,000 Ge'ez tokens merged into the SentencePiece model) after continued MLM pretraining.

Runs are bit-reproducible: enable_full_determinism, CUBLAS_WORKSPACE_CONFIG=:4096:8, dataloader_num_workers=0.

Usage

from transformers import AutoTokenizer, AutoModelForQuestionAnswering
import torch

repo = "Hailay/VEXMLM-TIGQA"
tokenizer = AutoTokenizer.from_pretrained(repo, subfolder="seed-42")
model = AutoModelForQuestionAnswering.from_pretrained(repo, subfolder="seed-42")
model.eval()

question = "ኤርትራ ኣብ ኣየናይ ክፍለ ዓለም ትርከብ?"
context  = "ኤርትራ ኣብ አፍሪቃ ክፍለ ዓለም እትርከብ ሃገር እያ።"
enc = tokenizer(question, context, return_tensors="pt", truncation=True, max_length=256)

with torch.no_grad():
    out = model(**enc)

start = out.start_logits.argmax()
end = out.end_logits.argmax()
print(tokenizer.decode(enc.input_ids[0][start:end + 1], skip_special_tokens=True))

Limitations

• Fine-tuned for Tigrinya on TIGQA only; performance on other languages, domains or label schemes is not characterised.
• The base model covers Amharic and Tigrinya; other Ge'ez-script languages were not part of pretraining.
• The pretraining text is of undocumented origin (licence unknown); samples contain religious translations alongside general web prose, and the model may reflect those distributions and any biases present in them.
• Single-configuration study: no hyperparameter search was performed, and baseline comparisons in the paper are single-seed.

Reproducibility

The fine-tuning launcher, evaluation code and per-run result records are in the official repository: https://github.com/hailaykidu/VEXMLM

sbatch scripts/slurm_stage2_spm_seeds.sh    # 6 tasks × 5 seeds
python3 evaluation/export_spm_results.py    # regenerates the metrics table

Citation

@inproceedings{teklehaymanot2026vexmlm,
  title     = {Vocabulary Expansion for Low-Resource African Languages:
               A Case Study in Amharic and Tigrinya},
  author    = {Teklehaymanot, Hailay Kidu and Yadeta, Debela Desalegn and
               Nejdl, Wolfgang},
  booktitle = {Proceedings of the Workshop on Language Models for
               Underserved Communities (LM4UC) at IJCAI},
  year      = {2026}
}

Accepted at the LM4UC Workshop, IJCAI 2026.

License

Apache 2.0, following xlm-roberta-base.