mlsanityData to 7 Oct 2026

SEA-CLIP-Tiny: Efficient Multilingual Text-Vision Embedding for Southeast Asian Languages

arXiv 2609.30739 ↗First posted 25 Sep 2026cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Multilingual text-vision embedding models are essential for cross-lingual image-text retrieval, but Southeast Asian languages remain poorly supported due to the region's linguistic diversity and limited data and computing resources. In this paper, we introduce SEA-CLIP-Tiny, a compact multilingual text-vision embedding model for Southeast Asia with fewer than 50M parameters. Our model adapts a CLIP-KD-style framework to Southeast Asian multilingual settings through regional data curation and multilingual teacher guidance. Experiments across seven Southeast Asian languages show that SEA-CLIP-Tiny achieves the strongest average retrieval performance among the evaluated student models, reaching 12.9%, 31.5%, and 42.2% at R@1, R@5, and R@10, respectively. Compared with MobileCLIP2, it improves average R@10 by 12.1 points while using 38.4% fewer parameters and lower measured CPU latency. These results highlight the importance of region-aware training for efficient multilingual text-vision models in Southeast Asia.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 24 of its 28 references that are in the map. First posted on arXiv 25 Sep 2026.

Builds on

24 of 28 references.

References · 24

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 2,534
  3. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyondcited by 2,179
  4. 2021Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Conceptscited by 1,209
  5. 2025SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Featurescited by 1,127
  6. 2021WIT: Wikipedia-based Image Text Dataset for Multimodal Multilingual Machine Learningcited by 285
  7. 2023Data Filtering Networkscited by 247
  8. 2024VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Taskscited by 222
  9. 2022Chinese CLIP: Contrastive Vision-Language Pretraining in Chinesecited by 108
  10. 2022Crossmodal-3600: A Massively Multilingual Multimodal Evaluation Datasetcited by 100
  11. 2023TinyCLIP: CLIP Distillation via Affinity Mimicking and Weight Inheritancecited by 90
  12. 2023CLIP-KD: An Empirical Study of CLIP Model Distillationcited by 87
12 more
  1. 2022AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilitiescited by 84
  2. 2025Meta CLIP 2: A Worldwide Scaling Recipecited by 59
  3. 2024All Languages Matter: Evaluating LMMs on Culturally Diverse 100 Languagescited by 55
  4. 2022Bloom Library: Multimodal Datasets in 300+ Languages for a Variety of Downstream Taskscited by 28
  5. 2025MobileCLIP2: Improving Multi-Modal Reinforced Trainingcited by 20
  6. 2021Contrastive Language-Image Pre-training for the Italian Languagecited by 19
  7. 2025SEA-HELM: Southeast Asian Holistic Evaluation of Language Modelscited by 17
  8. 2025Grounding Multilingual Multimodal LLMs With Cultural Knowledgecited by 16
  9. 2023Babel-ImageNet: Massively Multilingual Evaluation of Vision-and-Language Representationscited by 11
  10. 2025MIEB: Massive Image Embedding Benchmarkcited by 11
  11. 2026SEA-Guard: Culturally Grounded Multilingual Safeguard for Southeast Asiacited by 2
  12. 2026Anthropogenic Regional Adaptation in Multimodal Vision-Language Modelcited by 2

Built on by

Not cited yet.