def test_compile_simple_fp16(self):
compiled_1 = compiler.compile_source(source=_make_saved_model(), config=Config())
compiled_2 = compiler.compile_source(source=_make_saved_model(),
config=Config(optimization=True, supported_types=[tf.float16]))
self.assertLess(len(compiled_2.tflite_model), len(compiled_1.tflite_model))
def test_compile_simple(self):
compiled = compiler.compile_source(source=_make_saved_model(),
config=Config())
self.assertIsInstance(compiled.tflite_model, bytes)
self.assertEqual(compiled.input_formats,
[DataFormat.CHANNELS_FIRST, DataFormat.CHANNELS_LAST])
def test_compile_select_tf_op(self):
compiled = compiler.compile_source(
source=_make_saved_model(),
config=Config(supported_ops=[tf.lite.OpsSet.SELECT_TF_OPS]))
self.assertIsInstance(compiled.tflite_model, bytes)
self.assertEqual(compiled.input_formats,
[DataFormat.CHANNELS_FIRST, DataFormat.CHANNELS_LAST])
def test_invalid_data_type(self):
with self.assertRaises(ValueError):
Config.from_json({'supported_types': ['foobar']})
with self.assertRaises(ValueError):
Config.from_json({'supported_types': ['as_dtype']})
def test_from_env(self):
self.assertEqual(Config(optimization=False, supported_types=None), Config.from_env({}))
self.assertEqual(Config(optimization=False, supported_types=None), Config.from_env({'OPTIMIZATION': '0'}))
self.assertEqual(Config(optimization=True, supported_types=None), Config.from_env({'OPTIMIZATION': '1'}))
self.assertEqual(Config(optimization=False, supported_types=[tf.float16]),
Config.from_env({'SUPPORTED_TYPES': 'float16'}))
self.assertEqual(Config(optimization=False, supported_types=[tf.float16, tf.float32]),
Config.from_env({'SUPPORTED_TYPES': 'float16,float32'}))
self.assertEqual(Config(optimization=True, supported_types=[tf.float16, tf.float32]),
Config.from_env({'OPTIMIZATION': '1', 'SUPPORTED_TYPES': 'float16,float32'}))
self.assertEqual(Config(supported_ops=None), Config.from_env({}))
self.assertEqual(Config(supported_ops=[tf.lite.OpsSet.TFLITE_BUILTINS_INT8]),
Config.from_env({'SUPPORTED_OPS': 'TFLITE_BUILTINS_INT8'}))
self.assertEqual(Config(supported_ops=[tf.lite.OpsSet.SELECT_TF_OPS, tf.lite.OpsSet.TFLITE_BUILTINS_INT8]),
Config.from_env({'SUPPORTED_OPS': 'SELECT_TF_OPS,TFLITE_BUILTINS_INT8'}))
self.assertEqual(Config(inference_input_type=tf.float32), Config.from_env({}))
self.assertEqual(Config(inference_input_type=tf.float32),
Config.from_env({'INFERENCE_INPUT_TYPE': 'float32'}))
self.assertEqual(Config(inference_output_type=tf.float32), Config.from_env({}))
self.assertEqual(Config(inference_output_type=tf.float32),
Config.from_env({'INFERENCE_OUTPUT_TYPE': 'float32'}))
def test_from_json(self):
self.assertEqual(Config(optimization=False, supported_types=None), Config.from_json({}))
self.assertEqual(Config(optimization=False, supported_types=None), Config.from_json({'optimization': False}))
self.assertEqual(Config(optimization=True, supported_types=None), Config.from_json({'optimization': True}))
self.assertEqual(Config(optimization=False, supported_types=[tf.float16]),
Config.from_json({'supported_types': ['float16']}))
self.assertEqual(Config(optimization=False, supported_types=[tf.float16, tf.float32]),
Config.from_json({'supported_types': ['float16', 'float32']}))
self.assertEqual(Config(optimization=True, supported_types=[tf.float16, tf.float32]),
Config.from_json({'optimization': True, 'supported_types': ['float16', 'float32']}))
self.assertEqual(Config(supported_ops=None), Config.from_json({}))
self.assertEqual(Config(supported_ops=[tf.lite.OpsSet.TFLITE_BUILTINS_INT8]),
Config.from_json({'supported_ops': ['TFLITE_BUILTINS_INT8']}))
self.assertEqual(Config(supported_ops=[tf.lite.OpsSet.SELECT_TF_OPS, tf.lite.OpsSet.TFLITE_BUILTINS_INT8]),
Config.from_json({'supported_ops': ['SELECT_TF_OPS', 'TFLITE_BUILTINS_INT8']}))
self.assertEqual(Config(inference_input_type=tf.float32), Config.from_json({}))
self.assertEqual(Config(inference_input_type=tf.float32),
Config.from_json({'inference_input_type': 'float32'}))
self.assertEqual(Config(inference_output_type=tf.float32), Config.from_json({}))
self.assertEqual(Config(inference_output_type=tf.float32),
Config.from_json({'inference_output_type': 'float32'}))
def test_from_env(self):
self.assertEqual(Config(optimization=False, supported_types=None),
Config.from_env({}))
self.assertEqual(Config(optimization=False, supported_types=None),
Config.from_env({'OPTIMIZATION': '0'}))
self.assertEqual(Config(optimization=True, supported_types=None),
Config.from_env({'OPTIMIZATION': '1'}))
self.assertEqual(
Config(optimization=False, supported_types=[tf.float16]),
Config.from_env({'SUPPORTED_TYPES': 'float16'}))
self.assertEqual(
Config(optimization=False,
supported_types=[tf.float16, tf.float32]),
Config.from_env({'SUPPORTED_TYPES': 'float16,float32'}))
self.assertEqual(
Config(optimization=True, supported_types=[tf.float16,
tf.float32]),
Config.from_env({
'OPTIMIZATION': '1',
'SUPPORTED_TYPES': 'float16,float32'
}))
def test_from_json(self):
self.assertEqual(Config(optimization=False, supported_types=None),
Config.from_json({}))
self.assertEqual(Config(optimization=False, supported_types=None),
Config.from_json({'optimization': False}))
self.assertEqual(Config(optimization=True, supported_types=None),
Config.from_json({'optimization': True}))
self.assertEqual(
Config(optimization=False, supported_types=[tf.float16]),
Config.from_json({'supported_types': ['float16']}))
self.assertEqual(
Config(optimization=False,
supported_types=[tf.float16, tf.float32]),
Config.from_json({'supported_types': ['float16', 'float32']}))
self.assertEqual(
Config(optimization=True, supported_types=[tf.float16,
tf.float32]),
Config.from_json({
'optimization': True,
'supported_types': ['float16', 'float32']
}))