 dk
|
4e48fabf41
Remove 4 state paradigm
|
1 year ago |
 DESKTOP-4GKCI80\Neuracle
|
b791a4670c
修改运行参数
|
1 year ago |
|
DESKTOP-4GKCI80\Neuracle
|
0fe6c556f7
add side args in general grasping paradigm
|
1 year ago |
|
DESKTOP-4GKCI80\Neuracle
|
92bb35d4c2
Fix: for edge mode
|
1 year ago |
|
DESKTOP-4GKCI80\Neuracle
|
be4ba70c49
Fix: decision = 0, extend (for step mode)
|
1 year ago |
|
DESKTOP-4GKCI80\Neuracle
|
8d751c5a32
Fix: 反馈释放阶段修改为extend
|
1 year ago |
|
dk
|
9c17d931c6
Feat: support step mode and edge mode for hand controller
|
1 year ago |
|
dk
|
96def9ca3e
新版气动手在未完成抓握时可以进行下一步操作，取消原有的反应时间
|
1 year ago |
|
dk
|
4ed6c405e0
Fix: 无模型输入时模型设置为None
|
1 year ago |
|
dk
|
11019065bf
删除 '.pylintrc'
|
1 year ago |
|
dk
|
1348622c9b
删除 'CONTRIBUTING.md'
|
1 year ago |
|
dk
|
a0a19e193b
删除 '.pre-commit-config.yaml'
|
1 year ago |
|
dk
|
112c6f3c21
删除 'README.md'
|
1 year ago |
|
dk
|
9213e12150
删除已经无用的todo
|
1 year ago |
|
dk
|
37e6c73f64
Refactor: 统一使用settings中的路径设置，psychopy程序和离线脚本统一模型路径输入方式
|
1 year ago |
|
dk
|
f3fbb1c660
Revert: use 1202+04 data model
|
1 year ago |
|
dk
|
f11c3d0496
Feat: 使用气动手状态机模型，一次运动切换一次状态
|
1 year ago |
|
DESKTOP-4GKCI80\Neuracle
|
26590183e6
modify input parameters when calling hand_device.start func
|
1 year ago |
|
dk
|
aba3e026d4
Fix: 在线的重参考参数
|
1 year ago |
|
dk
|
fca7ca152f
Feat: add momentum for online (smooth p)
|
1 year ago |
|
dk
|
535b1392f5
Freegrasping范式适应2分类模式
|
1 year ago |
|
dk
|
b84e562e7c
Merge branch 'hmm-model' of dk/kraken into master
|
1 year ago |
|
dk
|
f1bab40d72
Fix: datagen 起始时刻需要大于步长
|
1 year ago |
|
dk
|
492a44b5d2
将event_to_stim移入utils
|
1 year ago |
|
dk
|
3412736b25
normalize confusion matrix
|
1 year ago |
|
dk
|
d33960ff7a
Refactor: 按照新架构更新hmm训练
|
1 year ago |
|
dk
|
838da7e2a9
Merge branch 'hmm-model' of dk/kraken into master
|
1 year ago |
|
dk
|
cc615a5c93
Refactor: 重构模型架构，将模型分解为3部分：feat_extractor (作用在连续数据), feat_embedder (作用在epoch数据), clf (分类). 并在validation和online中复用推理代码
|
1 year ago |
|
dk
|
ba3d7734ca
Refactor: 把模型保存移到utils里
|
1 year ago |
|
dk
|
a888a4bef7
Feat: 引入经验参数，估计运动的大概时间，默认范式时间为5s
|
1 year ago |
