README
Mobile-O Pre-Training Data
Cross-Modal Alignment · 9M Text-Image Pairs
📌 Overview
This dataset is used for Stage 1: Cross-Modal Alignment pre-training of Mobile-O, a unified multimodal model for on-device understanding and generation.
The goal of this stage is to align the DiT diffusion decoder and Mobile Conditioning Projector (MCP) with the frozen VLM backbone using large-scale text-image pairs.
📊 Dataset Composition
| Source | Samples | Description |
|---|---|---|
| JourneyDB | 4M | High-quality AI-generated images with captions |
| BLIP3o-Pretrain-Short-Caption | 5M | Each image paired with a short caption generated by Qwen/Qwen2.5-VL-7B-Instruct |
🏋️ Training Details
- Stage: 1 — Cross-Modal Alignment (Pre-training)
- Trainable components: DiT + Mobile Conditioning Projector (MCP)
- Frozen components: Visual encoders, LLM backbone, VAE
- Script:
pretrain.sh
🔗 Related Resources
| Resource | Link |
|---|---|
| 📄 Paper | arXiv |
| 💻 Code | GitHub |
| 🤗 SFT Data | Mobile-O-SFT |
| 🤗 Post-Training Data | Mobile-O-Post-Train |
| 🤗 Model (0.5B) | Mobile-O-0.5B |
| 🤗 Model (1.5B) | Mobile-O-1.5B |
📄 Citation
@article{shaker2026mobileo,
title={Mobile-O: Unified Multimodal Understanding and Generation on Mobile Device},
author={Shaker, Abdelrahman and Heakl, Ahmed and Muhammad, Jaseel and Thawkar, Ritesh and Thawakar, Omkar and Li, Senmao and Cholakkal, Hisham and Reid, Ian and Xing, Eric P. and Khan, Salman and Khan, Fahad Shahbaz},
journal={arXiv preprint arXiv:2602.20161},
year={2026}
}
🙏 Acknowledgments
We gratefully acknowledge the following datasets used in constructing this pre-training corpus:
Contributors
446 commits
Similar repos
README
Mobile-O Pre-Training Data
Cross-Modal Alignment · 9M Text-Image Pairs
📌 Overview
This dataset is used for Stage 1: Cross-Modal Alignment pre-training of Mobile-O, a unified multimodal model for on-device understanding and generation.
The goal of this stage is to align the DiT diffusion decoder and Mobile Conditioning Projector (MCP) with the frozen VLM backbone using large-scale text-image pairs.
📊 Dataset Composition
| Source | Samples | Description |
|---|---|---|
| JourneyDB | 4M | High-quality AI-generated images with captions |
| BLIP3o-Pretrain-Short-Caption | 5M | Each image paired with a short caption generated by Qwen/Qwen2.5-VL-7B-Instruct |
🏋️ Training Details
- Stage: 1 — Cross-Modal Alignment (Pre-training)
- Trainable components: DiT + Mobile Conditioning Projector (MCP)
- Frozen components: Visual encoders, LLM backbone, VAE
- Script:
pretrain.sh
🔗 Related Resources
| Resource | Link |
|---|---|
| 📄 Paper | arXiv |
| 💻 Code | GitHub |
| 🤗 SFT Data | Mobile-O-SFT |
| 🤗 Post-Training Data | Mobile-O-Post-Train |
| 🤗 Model (0.5B) | Mobile-O-0.5B |
| 🤗 Model (1.5B) | Mobile-O-1.5B |
📄 Citation
@article{shaker2026mobileo,
title={Mobile-O: Unified Multimodal Understanding and Generation on Mobile Device},
author={Shaker, Abdelrahman and Heakl, Ahmed and Muhammad, Jaseel and Thawkar, Ritesh and Thawakar, Omkar and Li, Senmao and Cholakkal, Hisham and Reid, Ian and Xing, Eric P. and Khan, Salman and Khan, Fahad Shahbaz},
journal={arXiv preprint arXiv:2602.20161},
year={2026}
}
🙏 Acknowledgments
We gratefully acknowledge the following datasets used in constructing this pre-training corpus:
Similar repos
Linked in READMEs
Contributors
446 commits