Upload folder using huggingface_hub

README.md

@@ -0,0 +1,63 @@
+---
+library_name: braindecode
+tags:
+- eeg
+- brain-signal-processing
+- foundation-model
+license: bsd-3-clause
+---
+
+# EEGPT Pretrained Model
+
+Pretrained EEGPT encoder weights for EEG signal processing from [EEGPT: Pretrained Transformer for Universal and Reliable Representation of EEG Signals](https://proceedings.neurips.cc/paper_files/paper/2024/file/4540d267eeec4e5dbd9dae9448f0b739-Paper-Conference.pdf).
+
+This model was pretrained on a large-scale EEG dataset using masked dual self-supervised learning.
+
+## Model Details
+
+- **Architecture:** Transformer-based encoder with 8 layers
+- **Embedding dimension:** 512
+- **Number of attention heads:** 8
+- **Patch size:** 64
+- **Parameters:** ~10M
+
+## Usage
+
+```python
+from braindecode.models import EEGPT
+import mne
+
+# Create channel info (example with 22 channels)
+ch_names = ["Fz", "FC3", "FC1", "FCz", "FC2", "FC4", "C5", "C3", "C1", "Cz",
+            "C2", "C4", "C6", "CP3", "CP1", "CPz", "CP2", "CP4", "P1", "Pz", "P2", "POz"]
+ch_types = ["eeg"] * len(ch_names)
+info = mne.create_info(ch_names=ch_names, sfreq=250., ch_types=ch_types)
+
+# Load pretrained model
+model = EEGPT.from_pretrained(
+    "braindecode/eegpt-pretrained",
+    n_outputs=4,  # number of classes for your task
+    n_chans=22,
+    n_times=1024,
+    chs_info=info["chs"],
+    sfreq=250.0
+)
+
+# Use for downstream tasks
+# model can now be fine-tuned on your specific EEG classification task
+```
+
+## Citation
+
+```bibtex
+@inproceedings{tang2024eegpt,
+  title={EEGPT: Pretrained Transformer for Universal and Reliable Representation of EEG Signals},
+  author={Tang, Xiwen and others},
+  booktitle={Advances in Neural Information Processing Systems},
+  year={2024}
+}
+```
+
+## License
+
+BSD-3-Clause (braindecode)

config.json

@@ -0,0 +1,18 @@
+{
+  "architectures": [
+    "EEGPT"
+  ],
+  "model_type": "eegpt",
+  "embed_dim": 512,
+  "embed_num": 4,
+  "depth": 8,
+  "num_heads": 8,
+  "patch_size": 64,
+  "patch_stride": 32,
+  "mlp_ratio": 4.0,
+  "qkv_bias": true,
+  "drop_prob": 0.0,
+  "attn_drop_rate": 0.0,
+  "drop_path_rate": 0.0,
+  "init_std": 0.02
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddff4d9a60f8afb7647537bc586c9b97fe18f4955b92ffb123b2dc868e4e7908
+size 101157984