Skip to content

Commit

Permalink
docs[readme]: adding links to weights
Browse files Browse the repository at this point in the history
  • Loading branch information
@marcelampc
marcelampc committed Sep 2, 2022
1 parent 7bfc0b1 commit 2832348
Showing 1 changed file with 33 additions and 19 deletions.
52 changes: 33 additions & 19 deletions README.MD
View file
Original file line number Diff line number Diff line change
Expand Up @@ -35,28 +35,38 @@ poe poe-torch-cuda11 # to install Pytorch with CUDA 11.6
```

## Data preparation
To generate out-of-focus dataset, you must download ... image from Cityscape and disparity maps.

To convert disparity maps to depth maps, we use:
To be added.

```shell
this to convert to depth maps
```
[//]: # (To generate out-of-focus dataset, you must download ... image from Cityscape and disparity maps.)

[//]: # ()
[//]: # (To convert disparity maps to depth maps, we use:)

[//]: # ()
[//]: # (```shell)

[//]: # (this to convert to depth maps)

[//]: # (```)

Then, these depth maps are used with this matlab/octave code to generate defocused images using the method... The original code was written by ... and ... in ...
[//]: # ()
[//]: # (Then, these depth maps are used with this matlab/octave code to generate defocused images using the method... The original code was written by ... and ... in ...)

run xyz to generate the dataset. Change the parameters according to the experiment in the paper, or personal usage.
[//]: # ()
[//]: # (run xyz to generate the dataset. Change the parameters according to the experiment in the paper, or personal usage.)

Then, define the following environment variable to the
[//]: # ()
[//]: # (Then, define the following environment variable to the )

## Model zoo

| name | url |
|--------------------| --- |
| infocus_color | [weights will be soon available][model-link] |
| infocus_gray | [weights will be soon available][model-link] |
| defocus_color | [weights will be soon available][model-link] |
| defocus_gray | [weights will be soon available][model-link] |
| name | url |
|--------------------|---------------------------------------------|
| infocus_color | [download_weight][infocus_color-model-link] |
| infocus_gray | [download_weight][infocus_gray-model-link] |
| defocus_color | [download_weight][defocus-color-link] |
| defocus_gray | [download_weight][defocus_gray-model-link] |

We suggest to save weights inside a folder called ```checkpoints/model_name```.

Expand All @@ -76,13 +86,13 @@ visdom -p $DISPLAY_PORT
To train a pretrained DeepLabV3 with one of our weights with a single GPU, run:

```shell
python main.py --train --name name-of-the-new-project-with-pretrained-weights --dataroot PATH_TO_CITYSCAPES --batchSize $batch_size --nEpochs $end --display_id $display_id --port $port --use_resize --data_augmentation f f --resume
python train_test_semseg.py --train --name name-of-the-new-project-with-pretrained-weights --dataroot PATH_TO_CITYSCAPES --batchSize $batch_size --nEpochs $end --display_id $display_id --port $port --use_resize --data_augmentation f f --resume
```

To train from scratch, run:

```shell
python main.py --train --name name-of-the-new-project --dataroot PATH_TO_CITYSCAPES --batchSize $batch_size --nEpochs $end --display_id $display_id --port $port --use_resize --data_augmentation f f
python train_test_semseg.py --train --name name-of-the-new-project --dataroot PATH_TO_CITYSCAPES --batchSize $batch_size --nEpochs $end --display_id $display_id --port $port --use_resize --data_augmentation f f
```


Expand All @@ -91,13 +101,13 @@ To train a pretrained DeepLabV3 with one of our weights with a single GPU, run:

Generate images only:
```shell
python main.py --test --test_only --name infocus_color --save_samples --use_resize --display_id 0 --dataroot PATH_TO_DATA
python train_test_semseg.py --test --test_only --name infocus_color --save_samples --use_resize --display_id 0 --dataroot PATH_TO_DATA
```


Generate images and evaluation:
```shell
python main.py --test --test_metrics --name infocus_color --use_resize --display_id 0 --dataroot ./datasets/public_datasets/Cityscapes
python train_test_semseg.py --test --test_metrics --name infocus_color --use_resize --display_id 0 --dataroot ./datasets/public_datasets/Cityscapes
```

As a result from this last run, you should get the outputs under the results file and the following metric results:
Expand All @@ -109,7 +119,7 @@ mIOU 0.6486886632431875

Generate only evaluation (resulting segmentation must be in the corresponding folder):
```shell
python main.py --test --evaluate_only --name infocus_color --use_resize --display_id 0 --dataroot ./datasets/public_datasets/Cityscapes
python train_test_semseg.py --test --evaluate_only --name infocus_color --use_resize --display_id 0 --dataroot ./datasets/public_datasets/Cityscapes
```

## License
Expand All @@ -123,3 +133,7 @@ Code (scripts) are released under the [MIT license][license].
[arxiv-paper]: https://arxiv.org/list/cs.CV/recent
[model-link]: broken
[license]: LICENSE
[infocus_color-model-link]: https://upciti-computer-vision-public.s3.eu-west-3.amazonaws.com/weights-privacy-aware-paper/infocus_color/0300.pth.tar
[infocus_gray-model-link]: https://upciti-computer-vision-public.s3.eu-west-3.amazonaws.com/weights-privacy-aware-paper/infocus_gray/0300.pth.tar
[defocus-color-link]: https://upciti-computer-vision-public.s3.eu-west-3.amazonaws.com/weights-privacy-aware-paper/defocus_color/0300.pth.tar
[defocus_gray-model-link]: https://upciti-computer-vision-public.s3.eu-west-3.amazonaws.com/weights-privacy-aware-paper/defocus_gray/0300.pth.tar

0 comments on commit 2832348

Please sign in to comment.