Model Testing throws "Features not generated yet" even though features have been generated already

Report bugs
#1

Question/Issue: Model testing fails with “Features are not generated yet”.

image
image799×399 34.4 KB

Project ID: 217622 (not public yet)

Context/Use case: Features have been generated already, and I’ve additionally tried re-generating all the features again under the regeneration tab.


Creating job... OK (ID: 8693807)

Scheduling job in cluster...
Job started
Generating features for MFE...
Not generating new features: features already generated and no options or files have changed.

No new features, skipping...

Generating features for MFE OK


Reducing dimensions for MFE...
No new features, skipping...

Reducing dimensions for MFE OK


Classifying data for Classifier...
Copying features from processing blocks...
Copying features from DSP block...
Copying features from DSP block OK
Copying features from processing blocks OK

Classifying data for int8 model...
Classifying data for float32 model...
Scheduling job in cluster...
Scheduling job in cluster...
Job started
Job started
Classifying data for Classifier OK


Features are not generated yet

Job failed (see above)
``
#2

Hello @Samleo,

I just checked your project and it seems you managed to fully train it.
Also I noticed that you have two DSP block but you’re using only one for your NN Classifier.

I’d remove the one you don’t use otherwise it will impact your deployed model performances.

If you want to compare the results between both DSP block, you can also use the EON tuner search spaces.

Best,