I’m trying to convert a TensorFlow model to Core ML format using coremltools, but I’m encountering an error during the conversion process. Here’s is the minimum reproducible example:

import numpy as np
import tensorflow as tf
import coremltools as ct

# Generate some dummy data
X_train = np.random.rand(100, 10)  # 100 samples, 10 features
y_train = np.random.rand(100, 1)   # 100 samples, 1 target

# Define a simple neural network model
model = tf.keras.Sequential([
    tf.keras.layers.Dense(64, activation='relu', input_shape=(10,)),
    tf.keras.layers.Dense(32, activation='relu'),
    tf.keras.layers.Dense(1, activation='linear')
])

# Compile the model
model.compile(optimizer='adam', loss='mse')

# Train the model
model.fit(X_train, y_train, epochs=10)

# Save the TensorFlow model
model.save('simple_nn_model.h5')

# Define the input shape and type for the Core ML model
input_shape = (10,)  # Match the input shape of the model
input_name = 'input'
output_name = 'output'

# Convert the TensorFlow model to Core ML format
coreml_model = ct.convert(model,
                           convert_to="mlprogram",
                           source='tensorflow')

# Save the Core ML model
coreml_model.save('SimpleNNModel.mlmodel')

However, when I run the conversion code, I get the following error:

AttributeError: 'Sequential' object has no attribute '_get_save_spec'

I’ve ensured that I’m using the latest versions of TensorFlow and coremltools, but the error persists. What could be causing this issue, and how can I resolve it?

Details:

TensorFlow version: 2.17.0
coremltools version: 7.2
Python version: 3.12

Any help would be greatly appreciated!