Having non-converted operations, even for simplest models

System information

Platform: Tried on Google Colab
TensorFlow version: 2.15.0

Steps to reproduce

Creating a Python file with the following content in Google Colab (let's call it test.py):
```
import tensorflow as tf
```

model = tf.keras.models.Sequential([ tf.keras.layers.InputLayer(input_shape=(1,)), tf.keras.layers.Dense(1) ])

converter = tf.lite.TFLiteConverter.from_keras_model(model) tflite_model = converter.convert()

with open('model.tflite', 'wb') as f: f.write(tflite_model)


- Calling it in Jupyter Notebook by `!python test.py`
- The output will be:

```Python
2024-01-28 11:17:11.939381: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-01-28 11:17:11.939450: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-01-28 11:17:11.941203: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2024-01-28 11:17:13.646897: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
2024-01-28 11:17:16.713671: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:378] Ignored output_format.
2024-01-28 11:17:16.713736: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:381] Ignored drop_control_dependency.
Summary on the non-converted ops:
---------------------------------
 * Accepted dialects: tfl, builtin, func
 * Non-Converted Ops: 1, Total Ops 6, % non-converted = 16.67 %
 * 1 ARITH ops

- arith.constant:    1 occurrences  (f32: 1)

  (f32: 1)

Note: running the code directly in Jupyter Notebook won't print anything

Problem

Why can it not convert all the operations in such a simple model?
Was it supposed to be like that?
Does having non-converted operations affect the performance of the model when deployed on a microcontroller using TFLM?
Is there a way to solve it?

google-ai-edge / LiteRT

Having non-converted operations, even for simplest models #84

System information

Steps to reproduce

Problem