CoCoT

non-profit
Activity Feed

AI & ML interests

None defined yet.

Recent Activity

sanaka87Β 
posted an update 11 months ago
view post
Post
3,760
πŸš€ Introducing VideoCoF: Unified Video Editing with a Temporal Reasoner (Chain-of-Frames)!

We’re excited to introduce VideoCoF, a unified framework for instruction-based video editing that enables temporal reasoning and ~4Γ— video length extrapolation, trained with only 50k video pairs. πŸ”₯

πŸ” What makes VideoCoF different?
🧠 Chain-of-Frames reasoning , mimic human thinking process like Seeing β†’ Reasoning β†’ Editing to apply edits accurately over time without external masks, ensuring physically plausible results.
πŸ“ˆ Strong length generalization β€” trained on 33-frame clips, yet supports multi-shot editing and long-video extrapolation (~4Γ—).
🎯 Unified fine-grained editing β€” Object Removal, Addition, Swap, and Local Style Transfer, with instance-level & part-level, spatial-aware control.

⚑ Fast inference update
πŸš€ H100: ~20s / video with 4-step inference, making high-quality video editing far more practical for real-world use.

πŸ”— Links
πŸ“„ Paper: https://arxiv.org/abs/2512.07469
πŸ’» Code: https://github.com/knightyxp/VideoCoF
πŸ€— Demo: XiangpengYang/VideoCoF
🧩 Models: XiangpengYang/VideoCoF
🌐 Project Page: https://videocof.github.io/

#VideoEditing #DiffusionModels #GenerativeAI #ComputerVision #AI
  • 2 replies
Β·
sanaka87Β 
posted an update about 1 year ago
view post
Post
3,729
Excited to share our Unified Multimodal Models new work Reconstruction Alignment (RecA)! πŸš€ Just 6 Γ— 80GB A100s Γ— 4.5 hours to boost BAGEL performance across all tasks! Outperforms FLUX-Kontext in image editing capabilities!

πŸ“„ Paper: https://alphaxiv.org/abs/2509.07295
πŸ’» Code: https://github.com/HorizonWind2004/reconstruction-alignment
πŸ€— HF Models: sanaka87/reca-68ad2176380355a3dcedc068
✍️ DEMO: sanaka87/BAGEL-RecA
🌐 Project Page: https://reconstruction-alignment.github.io
πŸ”₯ X: https://x.com/XDWang101/status/1965908302581420204
πŸ“° Zhihu: https://zhuanlan.zhihu.com/p/1947584568187159814
πŸ€— HF Daily Paper: Reconstruction Alignment Improves Unified Multimodal Models (2509.07295)

⚑ <10k images & 27 GPU hours (no-arch-changes) β†’ SOTA, surpassing much larger open-source & private models:

πŸ“Š GenEval: 0.73 β†’ 0.90 | πŸ“Š DPGBench: 80.93 β†’ 88.15
πŸ–ΌοΈ ImgEdit: 3.38 β†’ 3.75 | πŸ–ŒοΈ GEdit: 6.94 β†’ 7.25

βœ… RecA trains UMMs to reconstruct images from their own visual understanding encoder embeddings β†’ big gains in image generation 🎨 & editing βœ‚οΈ.
sanaka87Β 
posted an update over 1 year ago
view post
Post
3,177
sanaka87Β 
posted an update over 1 year ago
view post
Post
2,783
πŸš€ Excited to Share Our Latest Work: In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer~

🎨 Daily Paper:
In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (2504.20690)

πŸ”“ Code is now open source!
πŸ”₯ Huggingface DEMO: RiverZ/ICEdit
🌐 Project Website: https://river-zhang.github.io/ICEdit-gh-pages/
🏠 GitHub Repository: https://github.com/River-Zhang/ICEdit/blob/main/scripts/gradio_demo.py
πŸ€— Huggingface: sanaka87/ICEdit-MoE-LoRA
πŸ“„ arxiv Paper: In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (2504.20690)

πŸ”₯ Why it’s cool:
- Achieves high-quality, multi-task image editing.
- Uses only 1% of the training parameters and 0.1% of the training data compared to existing methods β€” extremely efficient
- Beats several commercial models on background preservation, ID control, and consistency
- Open-source, low-cost, faster, and stronger β€” think of it as the β€œDeepSeek of image editing” πŸ‘€

We also implemented a Gradio demo app, available directly in our GitHub repo! And we made a flashy demo video β€” happy to send it your way!
  • 1 reply
Β·
sanaka87Β 
posted an update almost 2 years ago
view post
Post
1,843
πŸš€ Excited to Share Our Latest Work: 3DIS & 3DIS-FLUX for Multi-Instance Layout-to-Image Generation! ❀️❀️❀️

🎨 Daily Paper: 3DIS-FLUX: simple and efficient multi-instance generation with DiT rendering (2501.05131)
πŸ”“ Code is now open source!
🌐 Project Website: https://limuloo.github.io/3DIS/
🏠 GitHub Repository: https://github.com/limuloo/3DIS
πŸ“„ 3DIS Paper: https://arxiv.org/abs/2410.12669
πŸ“„ 3DIS-FLUX Tech Report: https://arxiv.org/abs/2501.05131

πŸ”₯ Why 3DIS & 3DIS-FLUX?
Current SOTA multi-instance generation methods are typically adapter-based, requiring additional control modules trained on pre-trained models for layout and instance attribute control. However, with the emergence of more powerful models like FLUX and SD3.5, these methods demand constant retraining and extensive resources.

✨ Our Solution: 3DIS
We introduce a decoupled approach that only requires training a low-resolution Layout-to-Depth model to convert layouts into coarse-grained scene depth maps. Leveraging community and company pre-trained models like ControlNet + SAM2, we enable training-free controllable image generation on high-resolution models such as SDXL and FLUX.

🌟 Benefits of Our Decoupled Multi-Instance Generation:
1. Enhanced Control: By constructing scenes using depth maps in the first stage, the model focuses on coarse-grained scene layout, improving control over instance placement.
2. Flexibility & Preservation: The second stage employs training-free rendering methods, allowing seamless integration with various models (e.g., fine-tuned weights, LoRA) while maintaining the generative capabilities of pre-trained models.

Join us in advancing Layout-to-Image Generation! Follow and star our repository to stay updated! ⭐