Commit
•
9c774e3
1
Parent(s):
309c2e0
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,52 @@
|
|
1 |
---
|
2 |
license: mit
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: mit
|
3 |
---
|
4 |
+
# Model Card for pre-trained EEGNet models on mental imagery datasets
|
5 |
+
|
6 |
+
Collection of 12 neural networks trained for motor imagery decoding.
|
7 |
+
|
8 |
+
## Model Details
|
9 |
+
|
10 |
+
- **Architecture:** [EEGNetv4](https://braindecode.org/stable/generated/braindecode.models.EEGNetv4.html) by [Lawhern et. al (2018)](https://doi.org/10.1088/1741-2552/aace8c).
|
11 |
+
|
12 |
+
|
13 |
+
## How to Get Started with the Model
|
14 |
+
|
15 |
+
- **Download and load in memory:**
|
16 |
+
```python
|
17 |
+
import pickle
|
18 |
+
|
19 |
+
# download the model from the hub:
|
20 |
+
path_kwargs = hf_hub_download(
|
21 |
+
repo_id='PierreGtch/EEGNetv4',
|
22 |
+
filename='EEGNetv4_Lee2019_MI/kwargs.pkl',
|
23 |
+
)
|
24 |
+
path_params = hf_hub_download(
|
25 |
+
repo_id='PierreGtch/EEGNetv4',
|
26 |
+
filename='EEGNetv4_Lee2019_MI/model-params.pkl',
|
27 |
+
)
|
28 |
+
with open(path_kwargs, 'rb') as f:
|
29 |
+
kwargs = pickle.load(f)
|
30 |
+
module_cls = kwargs['module_cls']
|
31 |
+
module_kwargs = kwargs['module_kwargs']
|
32 |
+
|
33 |
+
# load the model with pre-trained weights:
|
34 |
+
torch_module = module_cls(**module_kwargs)
|
35 |
+
```
|
36 |
+
- **Details:** more details and potential use-case scenarios can be found in the notebook [here](https://neurotechlab.socsci.ru.nl/resources/pretrained_imagery_models/)
|
37 |
+
|
38 |
+
|
39 |
+
## Training Details
|
40 |
+
|
41 |
+
- **Training dataset:** Each model was trained on the dataset with corresponding name in the MOABB library (see [datasets list](https://neurotechx.github.io/moabb/dataset_summary.html#motor-imagery)).
|
42 |
+
- **Details:** For details on the training procedure, please refer to the poster [here](https://neurotechlab.socsci.ru.nl/resources/pretrained_imagery_models/).
|
43 |
+
|
44 |
+
## Evaluation
|
45 |
+
|
46 |
+
- **Cross-dataset transfer:** The transfer abilities of the models was tested on the same datasets as for training.
|
47 |
+
- **Details:** The results can be found on the poster [here](https://neurotechlab.socsci.ru.nl/resources/pretrained_imagery_models/).
|
48 |
+
|
49 |
+
## Model Card Authors
|
50 |
+
|
51 |
+
- **Modedels training and results by:** Pierre Guetschel
|
52 |
+
|