https://github.com/drsleep/nas-segm-pytorch
Code for Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells, CVPR '19
https://github.com/drsleep/nas-segm-pytorch
camvid cityscapes cvpr-2019 dense-per-pixel-tasks depth-estimation neural-architecture-search pytorch real-time semantic-segmentation wacv2020
Last synced: 6 months ago
JSON representation
Code for Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells, CVPR '19
- Host: GitHub
- URL: https://github.com/drsleep/nas-segm-pytorch
- Owner: DrSleep
- License: other
- Created: 2019-06-16T02:30:36.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2020-06-25T00:16:52.000Z (over 5 years ago)
- Last Synced: 2023-10-20T19:28:13.530Z (almost 2 years ago)
- Topics: camvid, cityscapes, cvpr-2019, dense-per-pixel-tasks, depth-estimation, neural-architecture-search, pytorch, real-time, semantic-segmentation, wacv2020
- Language: Smalltalk
- Size: 20.7 MB
- Stars: 147
- Watchers: 8
- Forks: 25
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Neural Architecture Search of Semantic Segmentation Models (in PyTorch)
[](https://paperswithcode.com/sota/monocular-depth-estimation-on-nyu-depth-v2?p=fast-neural-architecture-search-of-compact)
[](https://paperswithcode.com/sota/semantic-segmentation-on-pascal-voc-2012-val?p=fast-neural-architecture-search-of-compact)
[](https://paperswithcode.com/sota/semantic-segmentation-on-camvid?p=template-based-automatic-search-of-compact)
[](https://paperswithcode.com/sota/real-time-semantic-segmentation-on-cityscapes?p=template-based-automatic-search-of-compact)
[](https://paperswithcode.com/sota/semantic-segmentation-on-cityscapes-val?p=template-based-automatic-search-of-compact)
[](https://paperswithcode.com/sota/semantic-segmentation-on-cityscapes?p=template-based-automatic-search-of-compact)
This repository provides official models from two papers:
1. `Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells`, available [here](https://arxiv.org/abs/1810.10804);
2. `Template-Based Automatic Search of Compact Semantic Segmentation Architectures`, available [here](https://arxiv.org/abs/1904.02365).
For citations:
```
Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells
Vladimir Nekrasov, Hao Chen, Chunhua Shen, Ian Reid
CVPR, 2019
```
and
```
Template-Based Automatic Search of Compact Semantic Segmentation Architectures
Vladimir Nekrasov, Chunhua Shen, Ian Reid
WACV, 2020
```
## Updates
*22 May, 2020*: Added the search script for the WACV 2020 experiments on CityScapes.
*05 April, 2020* : Added decoder design and pre-trained segmentation models from the WACV 2020 paper ["Template-Based Automatic Search of Compact Semantic Segmentation Architectures"](https://arxiv.org/abs/1904.02365).
## Getting Started
For flawless reproduction of our results, the Ubuntu OS is recommended. The models have been tested using Python 3.6.
### Dependencies
```
pip3
Cython
cv2
jupyter-notebook
matplotlib
numpy
Pillow
torch>=1.0
torchvision
```
## Inference Examples
For the ease of reproduction, we have embedded all our examples inside Jupyter notebooks.
### CVPR 2019 Segmentation
Please refer to results on [PASCAL VOC](./examples/inference/VOC-segm.ipynb)
### CVPR 2019 Depth Estimation
Please refer to results on [NYUD-v2](./examples/inference/NYU-depth.ipynb)
### WACV 2020 Segmentation CityScapes
Please refer to results on [CityScapes](./examples/inference/WACV-CS-segm.ipynb)
### WACV 2020 Segmentation CamVid
Please refer to results on [CamVid](./examples/inference/WACV-CV-segm.ipynb)
## Note on the runtime of WACV 2020 models
In the paper we wrongly claimed that the latency of `arch0` and `arch1` on `2048x1048` inputs were `95.7±0.721` and `145±0.215`, correspondingly (on a single `1080Ti`). These numbers were incorrect as the model was in the training regime and not in the evaluation mode (`model.eval()` in PyTorch). Below are the characteristics of the discovered architectures with correct latency:
|| arch0 | arch1
| -------- |:-------------:| -----:|
|Number of parameters (19 output classes)|280147|268235
|Latency, ms on 2048x1024 inputs|52.25±0.03|97.11±0.24
|Latency, ms on 480x360 inputs|8.97±0.10|11.51±0.14
## Search
If you would like to search for architectures yourself, please follow the instructions below:
### CVPR 2019 (PASCAL VOC)
#### Prepare data
You would need to have PASCAL VOC segmentation dataset expanded with annotations from BSD.
After that, run in your terminal:
```bash
mkdir data/datasets
ln -s /path_to_voc/VOCdevkit data/datasets/
```
#### Running search
We rely on a Cython-based script for calculating mean IoU. In order to build the corresponding files, run the following:
```bash
cd src
python helpers/setup.py build_ext --build-lib=./helpers/
```
After that, you can execute `./examples/search/search.sh` that will start the search process.
***[!] Please note that all the hyperparameters were tuned for running the search process on 2 GPUs, each with at least 11GB of memory. In case when your setup differs, you would need to tune the hyperparameters accordingly.***
#### Checking the results
We output the log-file with all information on the search process. You can easily see top-performing architectures using the following command: `python src/helpers/num_uq.py path_to_genotypes.out`
### WACV 2020 (CityScapes)
#### Prepare data
You would need to have the CityScapes segmentation dataset.
After that, run in your terminal:
```bash
mkdir data/datasets
ln -s /path_to_cityscapes/cs/ data/datasets/
```
#### Running search
After that, you can execute `./examples/search/search_wacv.sh` that will start the search process.
***[!] Please note that all the hyperparameters were tuned for running the search process on 2 1080Ti GPUs. In case when your setup differs, you would need to tune the hyperparameters accordingly.***
## License
This project is free to use for non-commercial purposes - see the [LICENSE](LICENSE) file for details.
## Acknowledgments
* University of Adelaide and Australian Centre for Robotic Vision (ACRV) for making this project happen
* HPC Phoenix cluster at the University of Adelaide for making the training of the models possible
* PyTorch developers
* Yerba mate tea