Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.19324 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910153389899776 |
|---|---|
| author | Kerola, Tommi Masuda, Yuya Masuko, Takashi Nakanishi, Toshiki Nishino, Daisuke Takahashi, Kuniyuki Wang, Hanqin Yamada, Yoshihiro |
| author_facet | Kerola, Tommi Masuda, Yuya Masuko, Takashi Nakanishi, Toshiki Nishino, Daisuke Takahashi, Kuniyuki Wang, Hanqin Yamada, Yoshihiro |
| contents | We introduce PLaMo 2.1-VL, a lightweight Vision Language Model (VLM) for autonomous devices, available in 8B and 2B variants and designed for local and edge deployment with Japanese-language operation. Focusing on Visual Question Answering (VQA) and Visual Grounding as its core capabilities, we develop and evaluate the models for two real-world application scenarios: factory task analysis via tool recognition, and infrastructure anomaly detection. We also develop a large-scale synthetic data generation pipeline and comprehensive Japanese training and evaluation resources. PLaMo 2.1-VL outperforms comparable open models on Japanese and English benchmarks, achieving 61.5 ROUGE-L on JA-VG-VQA-500 and 85.2% accuracy on Japanese Ref-L4. For the two application scenarios, it achieves 53.9% zero-shot accuracy on factory task analysis, and fine-tuning on power plant data improves anomaly detection bbox + label F1-score from 39.7 to 64.9. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_19324 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PLaMo 2.1-VL Technical Report Kerola, Tommi Masuda, Yuya Masuko, Takashi Nakanishi, Toshiki Nishino, Daisuke Takahashi, Kuniyuki Wang, Hanqin Yamada, Yoshihiro Computer Vision and Pattern Recognition Artificial Intelligence We introduce PLaMo 2.1-VL, a lightweight Vision Language Model (VLM) for autonomous devices, available in 8B and 2B variants and designed for local and edge deployment with Japanese-language operation. Focusing on Visual Question Answering (VQA) and Visual Grounding as its core capabilities, we develop and evaluate the models for two real-world application scenarios: factory task analysis via tool recognition, and infrastructure anomaly detection. We also develop a large-scale synthetic data generation pipeline and comprehensive Japanese training and evaluation resources. PLaMo 2.1-VL outperforms comparable open models on Japanese and English benchmarks, achieving 61.5 ROUGE-L on JA-VG-VQA-500 and 85.2% accuracy on Japanese Ref-L4. For the two application scenarios, it achieves 53.9% zero-shot accuracy on factory task analysis, and fine-tuning on power plant data improves anomaly detection bbox + label F1-score from 39.7 to 64.9. |
| title | PLaMo 2.1-VL Technical Report |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2604.19324 |