GitLab is experiencing storage issues, linked with the Windows servers shutdown. Several features are affected by this disruption.

Check the info page for updates

Commit 4728548c authored by gencersumbul's avatar gencersumbul

Update README.md

parent 12b48203
......@@ -7,29 +7,29 @@ A paper describing the creation and evaluation of BigEarthNet-19 is currently un
We provide code and model weights for the following deep learning models that have been pre-trained on BigEarthNet with the new class nomenclature (BigEarthNet-19) for scene classification:
| Model Names | Pre-Trained TensorFlow Models | Pre-Trained PyTorch Models |
| ------------ | ------------------------------------------------------------ | ------------------------------------------------------------ |
| K-Branch CNN | [K-BranchCNN.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/K-BranchCNN.zip) | [K-BranchCNN.pth.tar](http://bigearth.net/static/pretrained-models-pytorch/BigEarthNet-19_labels/K-BranchCNN.pth.tar)|
| VGG16 | [VGG16.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/VGG16.zip) | [VGG16.pth.tar](http://bigearth.net/static/pretrained-models-pytorch/BigEarthNet-19_labels/VGG16.pth.tar) |
| VGG19 | [VGG19.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/VGG19.zip) | [VGG19.pth.tar](http://bigearth.net/static/pretrained-models-pytorch/BigEarthNet-19_labels/VGG19.pth.tar) |
| ResNet50 | [ResNet50.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/ResNet50.zip) | [ResNet50.pth.tar](http://bigearth.net/static/pretrained-models-pytorch/BigEarthNet-19_labels/ResNet50.pth.tar) |
| ResNet101 | [ResNet101.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/ResNet101.zip) | [ResNet101.pth.tar](http://bigearth.net/static/pretrained-models-pytorch/BigEarthNet-19_labels/ResNet101.pth.tar) |
| ResNet152 | [ResNet152.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/ResNet152.zip) | [ResNet152.pth.tar](http://bigearth.net/static/pretrained-models-pytorch/BigEarthNet-19_labels/ResNet152.pth.tar) |
| Model Names | Pre-Trained TensorFlow Models |
| ------------ | ------------------------------------------------------------ |
| K-Branch CNN | [K-BranchCNN.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/K-BranchCNN.zip) |
| VGG16 | [VGG16.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/VGG16.zip) |
| VGG19 | [VGG19.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/VGG19.zip) |
| ResNet50 | [ResNet50.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/ResNet50.zip) |
| ResNet101 | [ResNet101.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/ResNet101.zip) |
| ResNet152 | [ResNet152.zip](http://bigearth.net/static/pretrained-models/BigEarthNet-19_labels/ResNet152.zip) |
The TensorFlow code for these models can be found [here](https://gitlab.tu-berlin.de/rsim/bigearthnet-models-tf).
The PyTorch code for these models can be found [here](https://gitlab.tubit.tu-berlin.de/rsim/bigearthnet-models-pytorch).
The pre-trained models associated to other deep learning libraries will be released soon.
# Generation of Training/Test/Validation Splits
After downloading the raw images from https://www.bigearth.net, they need to be prepared for your ML application. We provide the script `prep_splits_BigEarthNet-19.py` for this purpose. It generates consumable data files (i.e., TFRecord) for training, validation and test splits which are suitable to use with TensorFlow or PyTorch. Suggested splits can be found with corresponding csv files under `splits` folder. The following command line arguments for `prep_splits_BigEarthNet-19.py` can be specified:
After downloading the raw images from https://www.bigearth.net, they need to be prepared for your ML application. We provide the script `prep_splits_BigEarthNet-19.py` for this purpose. It generates consumable data files (i.e., TFRecord) for training, validation and test splits which are suitable to use with TensorFlow. Suggested splits can be found with corresponding csv files under `splits` folder. The following command line arguments for `prep_splits_BigEarthNet-19.py` can be specified:
* `-r` or `--root_folder`: The root folder containing the raw images you have previously downloaded.
* `-o` or `--out_folder`: The output folder where the resulting files will be created.
* `-n` or `--splits`: A list of CSV files each of which contains the patch names of corresponding split.
* `-l` or `--library`: A flag to indicate for which ML library data files will be prepared: TensorFlow or PyTorch.
* `-l` or `--library`: A flag to indicate for which ML library data files will be prepared: TensorFlow.
* `--update_json`: A flag to indicate that this script will also change the original json files of the BigEarthNet by updating labels
To run the script, either the GDAL or the rasterio package should be installed. The TensorFlow package should also be installed. The script is tested with Python 2.7, TensorFlow 1.3, PyTorch 1.2 and Ubuntu 16.04.
To run the script, either the GDAL or the rasterio package should be installed. The TensorFlow package should also be installed. The script is tested with Python 2.7, TensorFlow 1.3, and Ubuntu 16.04.
**Note**: BigEarthNet patches with high density snow, cloud and cloud shadow are not included in the training, test and validation sets constructed by the provided scripts (see the list of patches with seasonal snow [here](http://bigearth.net/static/documents/patches_with_seasonal_snow.csv) and that of cloud and cloud shadow [here](http://bigearth.net/static/documents/patches_with_cloud_and_shadow.csv)).
......@@ -38,17 +38,8 @@ Authors
[**Gencer Sümbül**](http://www.user.tu-berlin.de/gencersumbul/)
**Jian Kang**
[**Tristan Kreuziger**](https://www.rsim.tu-berlin.de/menue/team/tristan_kreuziger/)
Maintained by
-------
[**Gencer Sümbül**](http://www.user.tu-berlin.de/gencersumbul/) for TensorFlow models
[**Mahdyar Ravanbakhsh**](https://www.rsim.tu-berlin.de/menue/team/dr_sayyed_mahdyar_ravanbakhsh/) for PyTorch models
# License
The BigEarthNet Archive is licensed under the **Community Data License Agreement – Permissive, Version 1.0** ([Text](https://cdla.io/permissive-1-0/)).
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment