INTERNAL ERROR: tuple index out of range

[DK.5]

I create a model in ONNX (in attach file)

I am unable to "Analyse" the model in X-Cube-AI Ver 5.1.2. 

All I get is an "INTERNAL ERROR: tuple index out of range" message

[fauvarque.daniel]

I've forwarded your model to our development team,

I'll keep you posted when we know the root cause and when it can be fixed

Regards

Daniel

[fauvarque.daniel]

Thanks again for reporting the issue, the problem has been reproduced in our latest release (5.2), it comes from an incorrect padding management in case of 1D tensors.

It will be fixed in X-Cube-AI 6.0 early next year.

Regards

Daniel

[YYin.2]

In version 7.0.0, I made this error again.

Can you tell me what I should do?

Thank you very much!

[fauvarque.daniel]

Your model is correctly supported now with X-CUBE-AI 7.3

Regards

Daniel

[fauvarque.daniel]

There is actually a bug in our code, normally fixed in the upcoming X-Cube-AI 7.1.0 early next year

Regards

Daniel

[lYC.1]

Hi,

In version 7.1.0,I created a model in ONNX.

I got the error: "INTERNAL ERROR: list index out of range".

Maybe it was the "onnx::pad" function that caused the error:

onnx::Pad[mode="constant", pads=[0, 0, -3, 0, 0, 0, -3, 0], value=0.]

I tried to comment out the pad function, and analyze completed successfully.

Here is the zip file that contains the onnx models before and after modification.

Can you tell me what I can do?

Thanks

[fauvarque.daniel]

This model is supported starting X-CUBE-AI version 7.2

Regards

Daniel

[fauvarque.daniel]

I've opened a new bug with your model so we can fix the import asap.

Thanks for reporting the issue

Regards

Daniel

[VVial.1]

Hi there,

I encounter the same problem with my onnx model... (INTERNAL ERROR: list index out of range)

Is there any solutions ?

Regards,

Valentin

[fauvarque.daniel]

Sorry for the late answer, this Robot model is correctly supported with X-CUBE-AI 7.3 already out

[EdgyAI]

I get the same problem now (January 2023) with a multi-input Keras model.

When analyzing the network, I see "INTERNAL ERROR: list index out of range".

The model has four inputs and is based on the model that won the 2016 PhysioNet challenge. ("Ensemble of Feature-based and Deep learning-based Classifiers for Detection of Abnormal Heart Sounds" by C. Potes, S. Parvaneh, A. Rahman, and B. Conroy).

Any word on workarounds or upcoming fixes?

[fauvarque.daniel]

After analysis the problem is likely to come from the usage of nested models.

Maybe you can try to rewrite the same topology with no or less nested levels

Regards

Daniel

[EdgyAI]

Daniel -

I finally got back to this. You were right. Once I removed the model nesting, the analysis worked correctly. I am glad I can even use multi-input Keras models.

Thank you!

[jonathanmeyer]

Hello all,

In version 8.1.0 Cube AI Cloud, I implemented a machine learning model by using Keras. 

I had an error and couldn't solve it.

INTERNAL ERROR: Unconsistent data. Filter size [2], strides (2,), in_shape(1, 240).

I checked my model and everything seems okey. My model has 3 inputs (X, Y, Z) and 6 output labels. I had constructed bunch of segments according to dataset. With this error text, which things should I control or change? 

Can you help me with that? Thank you.:.

 

Regards.

[pav65]

Facing Same issue now, Running XCube AI version 9.0.0 unable to analyze or validate my network.