metadata

language:
  - en
license: bsd-3-clause
tags:
  - arxiv:2506.23151
  - optical-flow-estimation
pipeline_tag: image-to-image
library_name: pytorch
base_model:
  - egorchistov/MEMFOF-Tartan-T

MEMFOF-Tartan-T-TSKH

📄 Paper | 🌐 Project Page | 💻 Code | 🚀 Colab | 🤗 Demo

🔍 This is a MEMFOF checkpoint trained from MEMFOF-Tartan-T on the combination of FlyingThings3D, Sintel, KITTI, and HD1K datasets.

✅ Note: This model is intended for real-world videos — it is trained with higher diversity and robustness in mind.

🛠️ Usage

git clone https://github.com/msu-video-group/memfof.git
cd memfof
pip3 install -r requirements.txt

import torch
from core.memfof import MEMFOF

device = "cuda" if torch.cuda.is_available() else "cpu"
model = MEMFOF.from_pretrained("egorchistov/MEMFOF-Tartan-T-TSKH").eval().to(device)

with torch.inference_mode():
    example_input = torch.randint(0, 256, [1, 3, 3, 1080, 1920], device=device)  # [B=1, T=3, C=3, H=1080, W=1920]
    backward_flow, forward_flow = model(example_input)["flow"][-1].unbind(dim=1)  # [B=1, C=2, H=1080, W=1920]

📚 Citation

@article{bargatin2025memfof,
  title={MEMFOF: High-Resolution Training for Memory-Efficient Multi-Frame Optical Flow Estimation},
  author={Bargatin, Vladislav and Chistov, Egor and Yakovenko, Alexander and Vatolin, Dmitriy},
  journal={arXiv preprint arXiv:2506.23151},
  year={2025}
}