Converter API change to enable optimization with sparsity.

To enable optimization with sparsity, use:
converter.optimizations = [tf.lite.Optimize.SPARSITY]
converter.convert()

Note:
1) This feature is experimental
2) It requires the use of during training pruning to be effective.
3) Not all kernels have been optimized for sparse execution, so the initial benefit will primarily be in the model size on disk.
PiperOrigin-RevId: 351245576

Author: liyunlu0618

tensorflow_model_optimization/python/examples/sparsity/keras/mnist/mnist_e2e.py

@@ -119,9 +119,8 @@ def main(unused_argv):
   with open(tflite_model_path, 'wb') as f:
     f.write(tflite_model_dense)
 
-  # The _experimental_sparsify_model is to enable sparse tensor encoding,
-  # otherwise the model is converted as dense.
-  converter._experimental_sparsify_model = True
+  # Enable sparse tensor encoding, otherwise the model is converted as dense.
+  converter.optimizations = {tf.lite.Optimize.EXPERIMENTAL_SPARSITY}
 
   tflite_model = converter.convert()
 
@@ -152,7 +151,7 @@ def main(unused_argv):
   model = train(model, x_train, y_train, x_test, y_test)
 
   converter = tf.lite.TFLiteConverter.from_keras_model(model)
-  converter._experimental_sparsify_model = True
+  converter.optimizations = {tf.lite.Optimize.EXPERIMENTAL_SPARSITY}
 
   tflite_model = converter.convert()
   # Check the model is compressed