Hi OpenViBE forum,
I concatenated two files into one using the Signal Concatenation Box. I also made sure to have one marker in the last file for the Classifier Trainer to initiate the training.
when I load the concatenated file into the mi-csp-3-classifier-trainer.xml scenario, the messaging box shows a lot of 'error' texts once the k-fold test is initiated. Please see the text below. Unforunately, I'm still new to OpenViBE and I don't understand what is causing the error
Any help to start moving in the right direction would be incredibly helpful! Thank you so much! - KevMat
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Received train stimulation. Data dim is [1650x6]
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> For information, we have 825 feature vector(s) for input 1
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> For information, we have 825 feature vector(s) for input 2
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Randomizing the feature vector set
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> k-fold test could take quite a long time, be patient
[ ERROR ] Plugin code caused crash 1 time(s)
[ ERROR ] [name:LDA Classifier]
[ ERROR] [class identifier:(0x2ba17a3c, 0x1bd46d84)]
[ ERROR ] [location:process callback]
[ ERROR ] <Algorithm::LDA Classifier> Classifier expected 0 features, got 6
...(repeated line of code removed)
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Finished with partition 1 / 5 (performance : 49.697%)
[ ERROR ] Plugin code caused crash 2 time(s)
...(repeated line of code removed)
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Finished with partition 2 / 5 (performance : 52.1212%)
[ ERROR ] Plugin code caused crash 3 time(s)
...(repeated line of code removed)
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Finished with partition 3 / 5 (performance : 50.6061%)
[ ERROR] Plugin code caused crash 4 time(s)
...(repeated line of code removed)
[ INF ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Finished with partition 4 / 5 (performance : 48.1818%)
[ ERROR ] Plugin code caused crash 5 time(s)
...(repeated line of code removed)
[ ERROR ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Training failed.
[ ERROR ] At time 609.250 sec <Box algorithm::(0x0a5a6a4a, 0x1d92a778) aka Classifier trainer> Bailing out (from whole set training)...
[WARNING] Box algorithm <Classifier trainer> has been deactivated because process() function returned error
Classifier Trainer - Linear Discriminant Analysis (LDA) - ERROR: Plugin code causing crash
Re: Classifier Trainer - Linear Discriminant Analysis (LDA) - ERROR: Plugin code causing crash
Hi, I have never seen this error. In case of, you use wich version of openvibe ? And a release or a personnal build Version.
Can you attach a folder with your scenario, scripts and data used.
All I can say is during the train, he test the classifier. I think, each time he try to train he cause crash, next, he test and say the error that the discriment function return a weight with a size of 0 instead of 6.
Can you click on trace button in log part to have more message.
Thibaut
Can you attach a folder with your scenario, scripts and data used.
All I can say is during the train, he test the classifier. I think, each time he try to train he cause crash, next, he test and say the error that the discriment function return a weight with a size of 0 instead of 6.
Can you click on trace button in log part to have more message.
Thibaut
Re: Classifier Trainer - Linear Discriminant Analysis (LDA) - ERROR: Plugin code causing crash
Hi Thibaut,
Thank you for looking into this!
I concatenate two trials into one data file (.gdf file format)
Both trial 1 and 2 are epoched and performed ICA (Independent Component Analysis) before combined, using EEGLAB.
I tried concatenate the two trials into one data file unedited but openvibe freezes
Version of OpenviBE: v1.2.2
mi-csp-2-train-CSP_KevMat.xml
**Headset: Customed EasyCap Emotiv+ headset
**Channels per frame: 14
**Frames per epoch: 640
**Epochs: 50
**Events: 50
**Sampling Rate (Hz): 128
**Epoch start (sec): -1.000
**Epoch End (sec): 3.992
**Reference: Nz
**Channel locations: Yes
I set the 'CSP Spatial Filter Trainer' Box to OVTK_StimulationID_EndOfFile train trigger. Filter is successfully trained
mi-csp-3-classifier-trainer_KevMat.xml
**same data, except no ICA performed
**No additional filtering
I set the 'Classifier trainer' to OVTK_StimulationID_EndOfFile and run scenario. This is when the LDA classifier fails.
Maybe it's because I am using different files for the GDF file reader? How can I combine two data trials combined successfully?
Best,
KevMat
Thank you for looking into this!
I concatenate two trials into one data file (.gdf file format)
Both trial 1 and 2 are epoched and performed ICA (Independent Component Analysis) before combined, using EEGLAB.
I tried concatenate the two trials into one data file unedited but openvibe freezes
Version of OpenviBE: v1.2.2
mi-csp-2-train-CSP_KevMat.xml
- Data Used for GDF File Reader: BRYAN-NO-SQUEEZE-micsp-trial 1 and 2 merged datasets - epochs pruned with ICA 769 770 only.gdf
**Headset: Customed EasyCap Emotiv+ headset
**Channels per frame: 14
**Frames per epoch: 640
**Epochs: 50
**Events: 50
**Sampling Rate (Hz): 128
**Epoch start (sec): -1.000
**Epoch End (sec): 3.992
**Reference: Nz
**Channel locations: Yes
I set the 'CSP Spatial Filter Trainer' Box to OVTK_StimulationID_EndOfFile train trigger. Filter is successfully trained
mi-csp-3-classifier-trainer_KevMat.xml
- Data Used for GDF File Reader: BRYAN-NO-SQUEEZE-micsp-trial12-merged_EPOCH.gdf
- Used trained csp filter in scenario 2
**same data, except no ICA performed
**No additional filtering
I set the 'Classifier trainer' to OVTK_StimulationID_EndOfFile and run scenario. This is when the LDA classifier fails.
Maybe it's because I am using different files for the GDF file reader? How can I combine two data trials combined successfully?
Best,
KevMat
- Attachments
-
- Classifier Trainer - Linear Discriminant Analysis (LDA) - ERROR Plugin Code Causing Crash.zip
- (12.65 KiB) Downloaded 173 times
Re: Classifier Trainer - Linear Discriminant Analysis (LDA) - ERROR: Plugin code causing crash
Also, how can I upload the data (.gdf) to this post? and, where's the trace button in the log part?
Thank you again!
Kevin Matsuno
Thank you again!
Kevin Matsuno
Re: Classifier Trainer - Linear Discriminant Analysis (LDA) - ERROR: Plugin code causing crash
HI, For the trace button see image
For the GDF maybe with an other site to temporaly share and suppress after
Thibaut
.For the GDF maybe with an other site to temporaly share and suppress after
Thibaut