Papers
arxiv:2611.16500

EvenVLM: A Text and Multimodal Dataset for the Even Language with Vision-Language Fine-Tuning

Published on Nov 12
· Submitted by
Daniil
on Nov 16
Authors:

Abstract

EvenVLM provides a text and multimodal VQA dataset for the endangered Even language, with tokenizer expansion and LoRA fine-tuning improving translation, text benchmarks, and VQA accuracy.

We present EvenVLM, the first text and multimodal (VQA) dataset for Even (Lamut), an endangered Northern Tungusic language spoken by approximately 6,800 people in Siberia. Our dataset comprises 50,521 structured Even-Russian entries extracted from dictionary and phrasebook sources via GPT-4o Mini-based parsing, from which we synthetically generate 107,543 training examples spanning translation, grammar, reading comprehension, and semantic classification. To address tokenization inefficiency on Even's agglutinative morphology, we propose a three-step tokenizer expansion algorithm that adds Even-specific graphemes (eng, oe, o-umlaut), whitelisted vocabulary, and BPE merges, reducing average tokens per word by 27% (4.99 to 3.64). Using LoRA fine-tuning of Qwen3-VL-4B, we adapt the model for text tasks (0.53% trainable parameters, rank 32) and separately for visual question answering (rank 64, frozen vision encoder). Our best configuration achieves 11.4% exact match and a 13.6 weighted score on text benchmarks, 0.1283 chrF for sentence translation, and 65% accuracy on VQA (up from 0% for the base model and text-only fine-tuning). We release our datasets, evaluation benchmarks, and training pipeline to support further research on Even and other low-resource languages.

Community

Paper author Paper submitter

We present EvenVLM, the first text and multimodal (VQA) dataset for Even, an endangered Northern Tungusic language spoken by ~6,800 people in Siberia. The dataset comprises 50,521 Even-Russian entries synthetically expanded to 107,543 training examples. We also propose a 3-step tokenizer expansion algorithm reducing token fragmentation by 27%, and fine-tune Qwen3-VL-4B using LoRA, achieving 65% VQA accuracy (up from 0% for the base model).

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2611.16500
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2611.16500 in a model README.md to link it from this page.

Datasets citing this paper 1

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.16500 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.