Issues with loading ML Program into MLModelAsset

Question

sami_elhini_roc OP

Created Jan ’25

Replies 1

Boosts 0

Participants 2

Good evening,

I am encountering an issue when trying to use this method:

modelAssetWithSpecificationData:blobMapping:error:

I provide the specification data and a blobMapping.

In the generated spec I can see the weights are referenced with this way:

@model_path/weights/weight.bin

In model.mlmodel.

I populate my blob mapping with the content of weights.bin with a URL that matches the above string.

However, I don't believe the blobMapping is ever being accessed because I get the following error:

Error: compiler error: Encountered an error while compiling a neural network model: in operation const_0_to_fp16: Cannot interpret @model_path when ModelPath is unset

I can confirm that the operation is the first use of weights in the model.mlmodel

I have not yet found a way to set ModelPath or to change that string.

Any assistance would be greatly appreciated.

Boost

Answer 1

DTS Engineer OP

Apple

3d

Thanks for the post and sorry for the delay.

Do you get the same results with just the relevant code in a small test project? If so, please share a link to your test project. That'll help us better understand what's going on. If you're not familiar with preparing a test project, take a look at Creating a test project.

Albert Pascual   Worldwide Developer Relations.

0