bruAristimunha commited on
Commit
6ffbaaf
·
verified ·
1 Parent(s): 07b24dc

Card: list the training data (datasets:, base_model:) and add a Training data section

Browse files

Follow-up to braindecode PR #1493: `from_pretrained` reads `datasets:` into `model.card_data`. Ids: MOABB names, NEMAR nm/on ids, OpenNeuro ds ids; evidence per checkpoint.

Files changed (1) hide show
  1. README.md +20 -0
README.md CHANGED
@@ -9,6 +9,8 @@ tags:
9
  - braindecode
10
  - emg2pose
11
  pipeline_tag: other
 
 
12
  ---
13
 
14
  # NeuroPose — emg2pose regression baseline
@@ -49,6 +51,24 @@ published baseline: encoder widths 32/128/256 pooling (10,2) (8,2) (4,4) over
49
  (time x electrodes), five residual blocks of three conv groups, decoder 128/32/1
50
  upsampling (10,4) (8,4) (4,2).
51
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
52
  ## License and attribution
53
 
54
  Weights are Meta's, released under **CC BY-NC-SA 4.0** — non-commercial, share-alike.
 
9
  - braindecode
10
  - emg2pose
11
  pipeline_tag: other
12
+ datasets:
13
+ - nm000281
14
  ---
15
 
16
  # NeuroPose — emg2pose regression baseline
 
51
  (time x electrodes), five residual blocks of three conv groups, decoder 128/32/1
52
  upsampling (10,4) (8,4) (4,2).
53
 
54
+ <!-- braindecode:training-data -->
55
+ ## Training data
56
+
57
+ The data that shaped these weights. The ids are the `datasets:` field of this card,
58
+ which braindecode reads into `model.card_data.datasets` when you call
59
+ `from_pretrained` (MOABB class names for MOABB datasets, OpenNeuro / NEMAR
60
+ accessions, Hub ids otherwise). `nm`/`on` ids are datasets on [NEMAR](https://nemar.org)
61
+ (`on` = NEMAR's copy of the OpenNeuro dataset with the same number). Rows marked
62
+ `private:` are data the authors did not release; they are not in `datasets:`.
63
+
64
+ | `datasets:` id | Dataset | Stage | Modality |
65
+ |---|---|---|---|
66
+ | [`nm000281`](https://nemar.org/dataset/nm000281) | [emg2pose: sEMG (sEMG-RD wristbands, 2 kHz, 16 ch/hand) + motion-capture hand pose, 193 users, 370 h (Salter et al. 2024, NeurIPS D&B)](https://github.com/facebookresearch/emg2pose) | training | EMG |
67
+
68
+ - experiment regression_neuropose (Liu et al. 2021 architecture re-trained by Meta on emg2pose; not Liu et al.'s own data). Supervised on the emg2pose train split (Table 3: 158 users, 23 stages, 250.9 h); val split monitored for checkpointing/early stopping (monitor val_loss); held-out user / stage / user-stage test sets not trained on.
69
+ - Sources: arXiv:2412.02725 (HTML) Table 3; https://github.com/facebookresearch/emg2pose README 'Downloading Pre-trained Checkpoints'; https://github.com/facebookresearch/emg2pose config/base.yaml.
70
+
71
+
72
  ## License and attribution
73
 
74
  Weights are Meta's, released under **CC BY-NC-SA 4.0** — non-commercial, share-alike.