A tool used to export monodepth2 models to ONNX file format.
Please refer to the original monodepth2 paper and github repository to find the original work:
Digging into Self-Supervised Monocular Depth Prediction
Clément Godard, Oisin Mac Aodha, Michael Firman and Gabriel J. Brostow
This code is for non-commercial use; please see the license file for terms.
Please follow the setup described in the original mondepth2 repository
Convert a monodepth2 model to onnx with:
python3 export_onnx.py --model_name=mono+stereo_640x192This will export the onnx file for mono+stereo_640x192. All models are exported with only the first output.
All pretrained monodepth2 models are supported
Available models:
- mono_640x192
- stereo_640x192
- mono+stereo_640x192
- mono_no_pt_640x192
- stereo_no_pt_640x192
- mono+stereo_no_pt_640x192
- mono_1024x320
- stereo_1024x320
- mono+stereo_1024x320
To use a custom model, refer to the convert_to_onnx() function.
Copyright © Niantic, Inc. 2019. Patent Pending. All rights reserved. Please see the license file for terms.