search

lzccccc / SMOKE

SMOKE: Single-Stage Monocular 3D Object Detection via Keypoint Estimation
MIT License
707 stars 177 forks source link

How to inference? #10

Closed lucasjinreal closed 4 years ago

lucasjinreal commented 4 years ago

Does there any inference code ?

lzccccc commented 4 years ago

I guess you mean "Is there any inference code?" The answer is Yes. Please check https://github.com/lzccccc/SMOKE/blob/master/tools/plain_train_net.py#L73

lzccccc commented 4 years ago

You can download the weights from Google drive: https://drive.google.com/file/d/11VK8_HfR7t0wm-6dCNP5KS3Vh-Qm686-/view.

And you can inference it with python tools/plain_train_net.py --eval-only --config-file "configs/smoke_gn_vector.yaml" It will generates .txt file on test set. The visualization code is not provided.

Steps before are all included in README file. Another suggestion for you is to first read the README carefully instead of opening an issue and questioning others' work.

lucasjinreal commented 4 years ago

If I opensource it, I will write a code for intuitively prediction and visualize. Also your readme doesn't provide more information about inference steps.

lzccccc commented 4 years ago

Different people have different ways of doing things. I already put everything including training/testing code and instructions in the repo. And there are others who could understand and use it easily. Use your own standard to require others is harsh...

anoop54 commented 4 years ago

This is fantastic as it is! Please ignore these comments, I think you've done an incredible job for the community.