locuslab / deq-flow Goto Github PK
View Code? Open in Web Editor NEW[CVPR 2022] Deep Equilibrium Optical Flow Estimation
License: GNU Affero General Public License v3.0
[CVPR 2022] Deep Equilibrium Optical Flow Estimation
License: GNU Affero General Public License v3.0
Hi,
Thanks for the great work.
Just wanted to know if you can also provide checkpoints for tiny model.
Thanks
Thanks for the great work.
The "fixed-point correction" appears to be applied in a dense manner, as seen in RAFT and similar methods. However, the paper mentions that it is applied in a sparse manner. What are the differences in the results between these two approaches?
Furthermore, it seems that intermediate hidden states need to be stored to compute the fixed-point correction. However, changing the f_thres did not result in any difference in memory usage. Can you explain why this is the case?
I want to use the pretrained model in my own images to predict optical flow, how can I do that?
I am new to optical flow. Thanks for your help.
Thanks for sharing this interesting work!
I have a question is that the results obtained by the model be converted into depth?
Hello, I am very interested in your work .I have read your paper"Deep Equilibrium Optical Flow Estimation",I want to know how many parameters the model DEQ-RAFT in Table 1 has.
deq-flow/code.v.2.0/core/utils/augmentor.py
Line 105 in 869d4bc
The crop augmentation here introduces a bug when the crop size is equal to the image size. Then numpy.random.randint gets a [0,0) range and throws an exception. This happened for me with some images equal in size to the crop size. The solution seems to me to simply add a +1.
Thank you for this interesting work.
Is there a demo that shows how I could run this over a video and get the optical flow per pair of frames?
can you please explain checkpoints training receipe to me?
as i read and understood,
deq-flow-H-chairs.pth === trained on flyingchairs
deq-flow-H-things-test-1x.pth == trained on 1FlyingChairs+1FlyingChairs
deq-flow-H-things-test-3x.pth == trained on 1FlyingChairs+3FlyingChairs
deq-flow-T-things-test.pth == no idea about this
deq-flow-B-chairs.pth == no idea
deq-flow-B-things-test.pth == no idea
deq-flow-H--chairs.pth == no idea
deq-flow-H-kitti.pth == no idea
deq-flow-H-sintel.pth == no idea
deq-flow-H-things-test-1.pth == no idea
deq-flow-H-things-test-2.pth == no idea
deq-flow-H-things-test-3.pth == no idea
and what is --eval_factor doing?
i didn't fina anything about eval_factor in repo
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.