Skip to content

Commit

Permalink
Add ramp/naip models
Browse files Browse the repository at this point in the history
  • Loading branch information
@bartoszptak
bartoszptak committed Feb 8, 2024
1 parent 24845ba commit 122466a
Show file tree
Hide file tree
Showing 2 changed files with 5 additions and 8 deletions.
2 changes: 1 addition & 1 deletion docs/source/creators/creators_example_onnx_model.rst
View file
Original file line number Diff line number Diff line change
Expand Up @@ -82,4 +82,4 @@ Steps based on the `tensorflow-onnx <https://github.com/onnx/tensorflow-onnx>`_
Update ONNX model to support dynamic batch size
===============================================

To convert model to support dynamic batch size, you need to update :code:`model.onnx` file. You can do it manually using `this <https://github.com/onnx/onnx/issues/2182#issuecomment-881752539>` script. Please note that the script is not perfect and may not work for all models.
To convert model to support dynamic batch size, you need to update :code:`model.onnx` file. You can do it manually using `this <https://github.com/onnx/onnx/issues/2182#issuecomment-881752539>`_ script. Please note that the script is not perfect and may not work for all models.
11 changes: 4 additions & 7 deletions docs/source/main/model_zoo/MODEL_ZOO.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,17 +2,17 @@

The [Model ZOO](https://chmura.put.poznan.pl/s/2pJk4izRurzQwu3) is a collection of pre-trained, deep learning models in the ONNX format. It allows for an easy-to-use start with the plugin.

NOTE: the provided models are not universal tools and will perform well only on similar data as in the training datasets. If you notice the model is not perfroming well on your data, consider re-training (or fine-tuning) it on your data.
> NOTE: the provided models are not universal tools and will perform well only on similar data as in the training datasets. If you notice the model is not perfroming well on your data, consider re-training (or fine-tuning) it on your data.
If you do not have machine learning expertise, feel free to contact the plugin authors for help or advice.
> If you do not have machine learning expertise, feel free to contact the plugin authors for help or advice.
## Segmentation models

| Model | Input size | CM/PX | Description | Example image |
|----------------------------------------------------------------------------------|------------|-------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------|
| [Corn Field Damage Segmentation](https://chmura.put.poznan.pl/s/abWFTVYSDIcncWs) | 512 | 3 | [PUT Vision](https://putvision.github.io/) model for Corn Field Damage Segmentation created on own dataset labeled by experts. We used the classical UNet++ model. It generates 3 outputs: healthy crop, damaged crop, and out-of-field area. | [Image](https://chmura.put.poznan.pl/s/i5WVmcfqPNdBTAQ) |
| [Land Cover Segmentation](https://chmura.put.poznan.pl/s/PnAFJw27uneROkV) | 512 | 40 | The model is trained on the [LandCover.ai dataset](https://landcover.ai.linuxpolska.com/). It provides satellite images with 25 cm/px and 50 cm/px resolution. Annotation masks for the following classes are provided for the images: building (1), woodland (2), water(3), road(4). We use `DeepLabV3+` model with `tu-semnasnet_100` backend and `FocalDice` as a loss function. NOTE: the dataset covers only the area of Poland, therefore the performance may be inferior in other parts of the world. | [Image](https://chmura.put.poznan.pl/s/Xa29vnieNQTvSt5) |
| [Buildings Segmentation](TODO) | 256 | 40 | Trained on the [RampDataset dataset](https://cmr.earthdata.nasa.gov/search/concepts/C2781412367-MLHUB.html). Annotation masks for buildings and background. Xunet network | [Image](ramp_example.png) |
| [Buildings Segmentation](https://chmura.put.poznan.pl/s/MwhgQNhyQF3fuBs) | 256 | 40 | Trained on the [RampDataset dataset](https://cmr.earthdata.nasa.gov/search/concepts/C2781412367-MLHUB.html). Annotation masks for buildings and background. Xunet network | [Image](https://chmura.put.poznan.pl/s/XCjuDKDS3FFovDl) |
| [Land Cover Segmentation Sentinel-2](TODO) | 64 | 1000 | Trained on the [Eurosat dataset](https://www.tensorflow.org/datasets/catalog/eurosat). Uses 13 spectral bands from Sentinel-2, with 10 classes. Model ConvNeXt. | [Image](eurosat_example.png) |
| [Agriculture segmentation RGB+NIR](TODO) | 256 | 30 | Trained on the Agriculture Vision 2021 dataset. 4 channels input (RGB + NIR). 9 output classes within agricultural field (weed_cluster, waterway, ...). Uses X-UNet. | [Image](agriculture_cision_2021_example.png) |
| [Fire risk assesment](TODO) | 384 | 1000 | Trained on the FireRisk dataset (RGB data). Classifies risk of fires (ver_high, high, low, ...). Uses ConvNeXt XXL. | [Image](firerisk.png) |
Expand All @@ -23,13 +23,12 @@ The [Model ZOO](https://chmura.put.poznan.pl/s/2pJk4izRurzQwu3) is a collection
| Model | Input size | CM/PX | Description | Example image |
|---------|---|---|---|---|
| | | | | |
| | | | | |

## Recognition models

| Model | Input size | CM/PX | Description | Example image |
|---------|---|---|---|---|
| NAIP Place recognition | 224 | 100 | ConvNeXt nano trained using SimSiam onn NAIP imagery | |
| [NAIP Place recognition](https://chmura.put.poznan.pl/s/k7EvbNGc2udHvck) | 224 | 100 | ConvNeXt nano trained using SimSiam onn [NAIP imagery](https://earth.esa.int/eogateway/catalog/pleiades-esa-archive) | [Image](https://chmura.put.poznan.pl/s/UzAvz8w5ceCui9y) |
| | | | | |

## Object detection models
Expand All @@ -48,8 +47,6 @@ The [Model ZOO](https://chmura.put.poznan.pl/s/2pJk4izRurzQwu3) is a collection
|[Residual Dense Network (RDN X4)](https://chmura.put.poznan.pl/s/AaKySmOoOhxW6qZ) |64 |Trained on 10 cm/px images set it same as input data | X4 | Model originally trained by H Zhang et. al. in "[A Comparative Study on CNN-Based Single-Image Super-Resolution Techniques for Satellite Images](https://github.com/farahmand-m/satellite-image-super-resolution)" converted to onnx format | [Image](https://chmura.put.poznan.pl/s/Ruz24ZpMNg97joV) from Massachusetts Roads Dataset [Dataset in kaggle](https://www.kaggle.com/datasets/balraj98/massachusetts-roads-dataset) |




## Contributing

PRs with models are welcome!
Expand Down

0 comments on commit 122466a

Please sign in to comment.