hackelle commited on
Commit
a799a75
1 Parent(s): 8d6d40e

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +105 -5
README.md CHANGED
@@ -1,9 +1,109 @@
1
  ---
 
2
  tags:
3
- - model_hub_mixin
4
- - pytorch_model_hub_mixin
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5
  ---
6
 
7
- This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
8
- - Library: [More Information Needed]
9
- - Docs: [More Information Needed]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ thumbnail: "https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/RSiM_Logo_1.png"
3
  tags:
4
+ - resnet50
5
+ - BigEarthNet v2.0
6
+ - Remote Sensing
7
+ - Classification
8
+ - image-classification
9
+ - Multispectral
10
+ library_name: configilm
11
+ license: mit
12
+ widget:
13
+ - src: example.png
14
+ example_title: Example
15
+ output:
16
+ - label: Agro-forestry areas
17
+ score: 0.000000
18
+ - label: Arable land
19
+ score: 0.000000
20
+ - label: Beaches, dunes, sands
21
+ score: 0.000000
22
+ - label: Broad-leaved forest
23
+ score: 0.000000
24
+ - label: Coastal wetlands
25
+ score: 0.000000
26
  ---
27
 
28
+ [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/)
29
+ :---:|:---:|:---:|:---:|:---:
30
+ <a href="https://www.tu.berlin/"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/tu-berlin-logo-long-red.svg" style="font-size: 1rem; height: 2em; width: auto" alt="TU Berlin Logo"/> | <a href="https://rsim.berlin/"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/RSiM_Logo_1.png" style="font-size: 1rem; height: 2em; width: auto" alt="RSiM Logo"> | <a href="https://www.dima.tu-berlin.de/menue/database_systems_and_information_management_group/"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/DIMA.png" style="font-size: 1rem; height: 2em; width: auto" alt="DIMA Logo"> | <a href="http://www.bigearth.eu/"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/BigEarth.png" style="font-size: 1rem; height: 2em; width: auto" alt="BigEarth Logo"> | <a href="https://bifold.berlin/"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/BIFOLD_Logo_farbig.png" style="font-size: 1rem; height: 2em; width: auto; margin-right: 1em" alt="BIFOLD Logo">
31
+
32
+ # Resnet50 pretained on BigEarthNet v2.0 using Sentinel-1 & Sentinel-2 bands
33
+
34
+ <!-- Optional images -->
35
+ <!--
36
+ [Sentinel-1](https://sentinel.esa.int/web/sentinel/missions/sentinel-1) | [Sentinel-2](https://sentinel.esa.int/web/sentinel/missions/sentinel-2)
37
+ :---:|:---:
38
+ <a href="https://sentinel.esa.int/web/sentinel/missions/sentinel-1"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/sentinel_2.jpg" style="font-size: 1rem; height: 10em; width: auto; margin-right: 1em" alt="Sentinel-2 Satellite"/> | <a href="https://sentinel.esa.int/web/sentinel/missions/sentinel-2"><img src="https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/sentinel_1.jpg" style="font-size: 1rem; height: 10em; width: auto; margin-right: 1em" alt="Sentinel-1 Satellite"/>
39
+ -->
40
+
41
+ This model was trained on the BigEarthNet v2.0 (also known as reBEN) dataset using the Sentinel-1 & Sentinel-2 bands.
42
+ It was trained using the following parameters:
43
+ - Number of epochs: up to 100 (with early stopping after 5 epochs of no improvement based on validation average
44
+ precision macro)
45
+ - Batch size: 512
46
+ - Learning rate: 0.001
47
+ - Dropout rate: 0.15
48
+ - Drop Path rate: 0.15
49
+ - Learning rate scheduler: LinearWarmupCosineAnnealing for 1000 warmup steps
50
+ - Optimizer: AdamW
51
+ - Seed: 0
52
+
53
+ The weights published in this model card were obtained after 19 training epochs.
54
+ 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.
55
+
56
+ ![[BigEarthNet](http://bigearth.net/)](https://raw.githubusercontent.com/wiki/lhackel-tub/ConfigILM/static/imgs/combined_2000_600_2020_0_wide.jpg)
57
+
58
+ The model was evaluated on the test set of the BigEarthNet v2.0 dataset with the following results:
59
+
60
+ | Metric | Macro | Micro |
61
+ |:------------------|------------------:|------------------:|
62
+ | Average Precision | 0.713316 | 0.862491 |
63
+ | F1 Score | 0.653815 | 0.764844 |
64
+ | Precision | 0.723217 | 0.806185 |
65
+
66
+ # Example
67
+ | A Sentinel-2 image (true color representation) |
68
+ |:---------------------------------------------------:|
69
+ | ![[BigEarthNet](http://bigearth.net/)](example.png) |
70
+
71
+ | Class labels | Predicted scores |
72
+ |:--------------------------------------------------------------------------|--------------------------------------------------------------------------:|
73
+ | <p> Agro-forestry areas <br> Arable land <br> Beaches, dunes, sands <br> ... <br> Urban fabric </p> | <p> 0.000000 <br> 0.000000 <br> 0.000000 <br> ... <br> 0.000000 </p> |
74
+
75
+
76
+ To use the model, download the codes that define the model architecture from the
77
+ [official BigEarthNet v2.0 (reBEN) repository](https://git.tu-berlin.de/rsim/reben-training-scripts) and load the model using the
78
+ code below. Note that you have to install [`configilm`](https://pypi.org/project/configilm/) to use the provided code.
79
+
80
+ ```python
81
+ from reben_publication.BigEarthNetv2_0_ImageClassifier import BigEarthNetv2_0_ImageClassifier
82
+
83
+ model = BigEarthNetv2_0_ImageClassifier.from_pretrained("path_to/huggingface_model_folder")
84
+ ```
85
+
86
+ e.g.
87
+
88
+ ```python
89
+ from reben_publication.BigEarthNetv2_0_ImageClassifier import BigEarthNetv2_0_ImageClassifier
90
+
91
+ model = BigEarthNetv2_0_ImageClassifier.from_pretrained(
92
+ "BIFOLD-BigEarthNetv2-0/resnet50-all-v0.1.1")
93
+ ```
94
+
95
+ If you use this model in your research or the provided code, please cite the following papers:
96
+ ```bibtex
97
+ CITATION FOR DATASET PAPER
98
+ ```
99
+ ```bibtex
100
+ @article{hackel2024configilm,
101
+ title={ConfigILM: A general purpose configurable library for combining image and language models for visual question answering},
102
+ author={Hackel, Leonard and Clasen, Kai Norman and Demir, Beg{\"u}m},
103
+ journal={SoftwareX},
104
+ volume={26},
105
+ pages={101731},
106
+ year={2024},
107
+ publisher={Elsevier}
108
+ }
109
+ ```