MEMFOF

MEMFOF is a memory-efficient multi-frame optical flow network from the MSU Graphics & Media Lab (ICCV 2025 Highlight). It reads three frames, t-1, t and t+1, and predicts the flow of the middle frame, aimed at Full HD video. It is the one flow family in LibreYOLO that reads a three-frame window.

Tasks
optical flow
Sizes
m at 1920 px
Install
pip install libreyolo
Support tier
Inference only, since v1.8.0. Predict, validate and export only. Training features do not apply.
Upstream
MEMFOF by MSU Graphics & Media Lab, BSD-3-Clause. Paper, source
Licenses
Code MIT, weights BSD-3-Clause. Commercial use

MEMFOF is a RAFT descendant that predicts the backward (t to t-1) and forward (t to t+1) flow of the middle frame in one pass. LibreYOLO reports the forward flow as Results.flow; the backward flow is computed but kept out of the task contract for now. The runner hands it the (t-1, t, t+1) window and repeats frame 0 as t-1 at the start of a sequence, so you still pass an ordinary list, folder or video. Pick it for real-world video at high resolution.

Install

MEMFOF needs no optional extra. Everything it imports is in the base install.

bash
pip install libreyolo

Predict

Weights download from the LibreYOLO Hugging Face organization on first use. The class LibreMEMFOF is also exported from libreyolo if you prefer to build the model directly.

Python
from libreyolo import LibreYOLO model = LibreYOLO("LibreMEMFOFm-flow.pt")   # weights download on first useresults = model(["frame_0001.jpg", "frame_0002.jpg"]) flow = results[0].flowprint(flow.data.shape)        # (H, W, 2) float32, (dx, dy) in pixelsprint(flow.mean_magnitude)    # mean displacement in pixels
Video
from libreyolo import LibreYOLO model = LibreYOLO("LibreMEMFOFm-flow.pt") # N frames give N - 1 results; result i is the flow from frame i to i + 1.for result in model("video.mp4", stream=True, save=True, save_flow=True):    print(result.flow.max_magnitude)
CLI
libreyolo predict model=LibreMEMFOFm-flow.pt source=frame_0001.jpg,frame_0002.jpg save=truelibreyolo predict model=LibreMEMFOFm-flow.pt source=video.mp4 save=true arrows=true save_flow=true

Flow needs at least two frames: a single image raises. A list of images, a directory sorted by name, a video or a stream is one sequence, and N frames give N - 1 results. result.flow is a FlowField with an (H, W, 2) float32 array on the original canvas of the first frame; flow.data[y, x] is (dx, dy) in pixels, +x right and +y down. MEMFOF fills no confidence map, so flow.confidence is None. result.boxes is None. save=True writes the Middlebury colour wheel, arrows=True or warp=True pick the other views, and save_flow=True also writes each pair's raw field as a .flo file. The optical flow task page describes the payload and views in full.

Variants

One architecture (ResNet-34 FPN), size m, with six checkpoints:

FileUpstream checkpoint
LibreMEMFOFm-flow.ptMEMFOF-Tartan-T-TSKH, recommended by the authors for real-world video (canonical)
LibreMEMFOFm-flow-tartan-t.ptan earlier published training stage
LibreMEMFOFm-flow-tartan.ptan earlier published training stage
LibreMEMFOFm-flow-kitti.ptKITTI fine-tune
LibreMEMFOFm-flow-sintel.ptSintel fine-tune
LibreMEMFOFm-flow-spring.ptSpring fine-tune

Frames run at native resolution with upstream's replicate padding to a multiple of 16. Frames whose long side exceeds imgsz (default 1920, MEMFOF's Full HD target) are downscaled first and the field is scaled back. Inference is fp32 with 8 refinement iterations, upstream's defaults.

Training is not offered: train() raises NotImplementedError.

Validate

val() runs the shared flow validator on a flow dataset (frames per sequence folder, Middlebury .flo, KITTI 16-bit PNG or .npy ground truth; see the dataset format). It reports the mean end-point error metrics/epe, metrics/px1 and metrics/px3 (share of valid pixels under 1 and 3 px of error) and the KITTI outlier rate metrics/fl_all.

Python
from libreyolo import LibreYOLO model = LibreYOLO("LibreMEMFOFm-flow.pt")metrics = model.val(data="my_flow.yaml") print(metrics["metrics/epe"])      # mean end-point error, pixelsprint(metrics["metrics/px1"])      # share of pixels under 1 px errorprint(metrics["metrics/fl_all"])   # KITTI outlier rate
CLI
libreyolo val model=LibreMEMFOFm-flow.pt data=my_flow.yaml

Export

TaskONNXTorchScriptExecuTorchTensorRTOpenVINOPaddleMNNRKNNAxelerancnnTFLiteCoreMLCore AI
Optical flowOptical flow to ONNX: not supportedOptical flow to TorchScript: not supportedOptical flow to ExecuTorch: not supportedOptical flow to TensorRT: not supportedOptical flow to OpenVINO: not supportedOptical flow to Paddle: not supportedOptical flow to MNN: not supportedOptical flow to RKNN: not supportedOptical flow to Axelera: not supportedOptical flow to ncnn: not supportedOptical flow to TFLite: not supportedOptical flow to CoreML: not supportedOptical flow to Core AI: not supported

Export is not available yet for any format. Optical flow export is blocked until the two-input graph (image1, image2) and the backend vector-rescale contract of the flow task are implemented for a family.

Checkpoints

Every published weight file for this family.

Every file above exists in the LibreYOLO org today and downloads on first use.

Licensing

Check the license on the Hugging Face repository of the specific weights you download. Every checkpoint in the LibreYOLO org carries one, and they are not always the same across a family. That repository is the authoritative source; the summary below describes what applied when this page was last verified.

This is a description of the licenses involved, not legal advice. If the answer matters commercially, read the licenses yourself and take your own counsel.

Original work
MEMFOF, MSU Graphics & Media Lab
Upstream license
BSD-3-Clause
LibreYOLO code
MIT
Weights
BSD-3-Clause, republished at huggingface.co/LibreYOLO
Interpretation
BSD-3-Clause is a permissive license: the code and weights can be used in commercial and closed-source products. Redistributions must keep the copyright notice, the license conditions and the disclaimer, and the authors' names may not be used to endorse derived products without permission. The upstream inference code at commit a51de9fc is vendored under libreyolo/models/memfof/_vendor/ with its BSD-3-Clause LICENSE; its gma.py attention modules come from GMA (zacjiang/GMA) under the WTFPL, shipped as LICENSE-GMA. LibreYOLO's wrapper code is MIT. The upstream checkpoints on Hugging Face (egorchistov/optical-flow-MEMFOF-*) are tagged bsd-3-clause but ship no license file, so LibreYOLO mirrors the six converted checkpoints (canonical, kitti, sintel, spring, tartan, tartan-t) under the code repository's BSD-3-Clause license. No gating; inference only.

MEMFOF is the work of the MSU Graphics & Media Lab. The upstream inference code is vendored under its BSD-3-Clause licence; its attention modules come from GMA under the WTFPL. The upstream checkpoints on Hugging Face are tagged bsd-3-clause but ship no licence file, so LibreYOLO mirrors the six converted checkpoints under the code repository's BSD-3-Clause licence, without gating. BSD-3-Clause allows commercial and closed-source use as long as the copyright notice, conditions and disclaimer are kept.

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}
}

Copied from the authors' citation block at github.com/msu-video-group/memfof.

Verified against LibreYOLO v1.8.0. Support tables, checkpoints and benchmark numbers on this page are generated from the released library and the published weights, not written by hand.