how to convert imagenet/mobilenet_v2_130_224/classification into mlmodel

Question

i m very new to this but I am trying to solve the issue somehow. this is my convert.py file

import numpy as n
import tensorflow as tf
import coremltools as ct
print(n.__version__)
print(tf.__version__)
print(ct.__version__)

loaded_model = tf.saved_model.load("mobilenet_v2_130_224.h5")
mlmodel = ct.convert(loaded_model, inputs=[ct.ImageType()], classifier_config=ct.ClassifierConfig("labels.txt"), source='tensorflow')
mlmodel.short_description = "My Classifier"
mlmodel.license = "Apache 2.0"
spec = mlmodel.get_spec()
ct.utils.rename_feature(spec, "input_1", "imageInput")
ct.utils.rename_feature(spec, "Identity", "classResult")
mlmodel = ct.models.MLModel(spec)
print(mlmodel)
mlmodel.save("model_299x299.mlmodel")

I downloaded the model from [here][1] and unzipped on desktop. I have M1 iMac. why do I get these errors below? how can I convert this particular model into mlmodel? if my convert file is wrongly written what should be the best practice of it ?

Traceback (most recent call last):
  File "/Users/asduskun/Desktop/convert.py", line 20, in <module>
    mlmodel = ct.convert(loaded_model, inputs=[ct.ImageType()], classifier_config=ct.ClassifierConfig("labels.txt"), source='tensorflow')
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/_converters_entry.py", line 444, in convert
    mlmodel = mil_convert(
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/converter.py", line 187, in mil_convert
    return _mil_convert(model, convert_from, convert_to, ConverterRegistry, MLModel, compute_units, **kwargs)
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/converter.py", line 211, in _mil_convert
    proto, mil_program = mil_convert_to_proto(
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/converter.py", line 281, in mil_convert_to_proto
    prog = frontend_converter(model, **kwargs)
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/converter.py", line 99, in __call__
    return tf2_loader.load()
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/frontend/tensorflow/load.py", line 61, in load
    self._graph_def = self._graph_def_from_model(output_names)
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/frontend/tensorflow2/load.py", line 133, in _graph_def_from_model
    cfs, graph_def = self._get_concrete_functions_and_graph_def()
  File "/Users/asduskun/miniconda3/lib/python3.10/site-packages/coremltools/converters/mil/frontend/tensorflow2/load.py", line 125, in _get_concrete_functions_and_graph_def
    raise NotImplementedError(msg.format(self.model))
NotImplementedError: Expected model format: [SavedModel | [concrete_function] | tf.keras.Model | .h5 | GraphDef], got <tensorflow.python.saved_model.load.Loader._recreate_base_user_object.<locals>._UserObject object at 0x1676455a0>

  [1]: https://tfhub.dev/google/imagenet/mobilenet_v2_130_224/classification/5

Answer 1

The error pops out of the ct.convert(ml_model) function. The model format comes from the saved model of tensorflow.

The most convenient way to convert from TensorFlow 2 is to use an object of the tf.keras.Model class.

If you download a pre-trained model (SavedModel or HDF5), first check that you can load it as a tf.keras.Model and run the predict() method on it.

Then pass the model into the Core ML Tools converter.

The recommended code for your reference with MobileNet is as follows :

import tensorflow as tf
import tensorflow_hub as tf_hub
import numpy as np

model = tf.keras.Sequential([
        tf.keras.layers.InputLayer(input_shape=(192, 192, 3)),
        tf_hub.KerasLayer(
          "https://tfhub.dev/google/imagenet/mobilenet_v2_050_192/classification/4"
        )
])

model.build([1, 192, 192, 3])  # Batch input shape.

Before this, you can do a !pip install tensorflow==2.6.2

# random input data to check that predict works
x = np.random.rand(1, 192, 192, 3)
tf_out = model.predict([x])

import coremltools as ct

# convert to Core ML and check predictions
mlmodel = ct.convert(model, convert_to="mlprogram")

Finally,

coreml_out_dict = mlmodel.predict({"image":x})
coreml_out = list(coreml_out_dict.values())[0]
np.testing.assert_allclose(tf_out, coreml_out, rtol=1e-2, atol=1e-1)

# convert to an image input Core ML model
# mobilenet model expects images to be normalized in the interval [-1,1]
# hence bias of -1 and scale of 1/127
mlmodel = ct.convert(model, convert_to="mlprogram",
                    inputs=[ct.ImageType(bias=[-1,-1,-1], scale=1/127)])

mlmodel.save("mobilenet.mlpackage")