--- thumbnail: "https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/RSiM_Logo_1.png" tags: - convnextv2_base - BigEarthNet v2.0 - Remote Sensing - Classification - image-classification - Multispectral library_name: configilm license: mit widget: - src: example.png example_title: Example output: - label: Agro-forestry areas score: 0.000005 - label: Arable land score: 0.000090 - label: Beaches, dunes, sands score: 0.000094 - label: Broad-leaved forest score: 0.000325 - label: Coastal wetlands score: 0.000057 --- [TU Berlin](https://www.tu.berlin/) | [RSiM](https://rsim.berlin/) | [DIMA](https://www.dima.tu-berlin.de/menue/database_systems_and_information_management_group/) | [BigEarth](http://www.bigearth.eu/) | [BIFOLD](https://bifold.berlin/) :---:|:---:|:---:|:---:|:---: TU Berlin Logo | RSiM Logo | DIMA Logo | BigEarth Logo | BIFOLD Logo # Convnextv2_base pretrained on BigEarthNet v2.0 using Sentinel-1 bands This model was trained on the BigEarthNet v2.0 (also known as reBEN) dataset using the Sentinel-1 bands. It was trained using the following parameters: - Number of epochs: up to 100 (with early stopping after 5 epochs of no improvement based on validation average precision macro) - Batch size: 512 - Learning rate: 0.001 - Dropout rate: 0.15 - Drop Path rate: 0.15 - Learning rate scheduler: LinearWarmupCosineAnnealing for 1000 warmup steps - Optimizer: AdamW - Seed: 42 The weights published in this model card were obtained after 18 training epochs. For more information, please visit the [official BigEarthNet v2.0 (reBEN) repository](https://git.tu-berlin.de/rsim/reben-training-scripts), where you can find the training scripts. ![[BigEarthNet](http://bigearth.net/)](https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/combined_2000_600_2020_0_wide.jpg) The model was evaluated on the test set of the BigEarthNet v2.0 dataset with the following results: | Metric | Macro | Micro | |:------------------|------------------:|------------------:| | Average Precision | 0.602211 | 0.789338 | | F1 Score | 0.548913 | 0.696168 | | Precision | 0.602211 | 0.789338 | # Example | A Sentinel-1 image (VV, VH and VV/VH bands are used for visualization) | |:---------------------------------------------------:| | ![[BigEarthNet](http://bigearth.net/)](example.png) | | Class labels | Predicted scores | |:--------------------------------------------------------------------------|--------------------------------------------------------------------------:| |

Agro-forestry areas
Arable land
Beaches, dunes, sands
...
Urban fabric

|

0.000005
0.000090
0.000094
...
0.000058

| To use the model, download the codes that define the model architecture from the [official BigEarthNet v2.0 (reBEN) repository](https://git.tu-berlin.de/rsim/reben-training-scripts) and load the model using the code below. Note that you have to install [`configilm`](https://pypi.org/project/configilm/) to use the provided code. ```python from reben_publication.BigEarthNetv2_0_ImageClassifier import BigEarthNetv2_0_ImageClassifier model = BigEarthNetv2_0_ImageClassifier.from_pretrained("path_to/huggingface_model_folder") ``` e.g. ```python from reben_publication.BigEarthNetv2_0_ImageClassifier import BigEarthNetv2_0_ImageClassifier model = BigEarthNetv2_0_ImageClassifier.from_pretrained( "BIFOLD-BigEarthNetv2-0/convnextv2_base-s1-v0.1.1") ``` If you use this model in your research or the provided code, please cite the following papers: ```bibtex CITATION FOR DATASET PAPER ``` ```bibtex @article{hackel2024configilm, title={ConfigILM: A general purpose configurable library for combining image and language models for visual question answering}, author={Hackel, Leonard and Clasen, Kai Norman and Demir, Beg{\"u}m}, journal={SoftwareX}, volume={26}, pages={101731}, year={2024}, publisher={Elsevier} } ```