def testGraphDefQuantizationInvalid(self):
in_tensor_1 = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="inputA")
in_tensor_2 = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="inputB")
_ = array_ops.fake_quant_with_min_max_args(in_tensor_1 + in_tensor_2,
min=0.,
max=1.,
name="output")
sess = session.Session()
input_arrays_map = [("inputA", [1, 16, 16, 3]),
("inputB", [1, 16, 16, 3])]
output_arrays = ["output"]
with self.assertRaises(ValueError) as error:
convert.toco_convert_graph_def(
sess.graph_def,
input_arrays_map,
output_arrays,
inference_type=lite_constants.QUANTIZED_UINT8)
self.assertEqual(
"std_dev and mean must be defined when inference_input_type is "
"QUANTIZED_UINT8.", str(error.exception))
def testGraphDefQuantizationInvalid(self):
with ops.Graph().as_default():
in_tensor_1 = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="inputA")
in_tensor_2 = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="inputB")
_ = array_ops.fake_quant_with_min_max_args(in_tensor_1 +
in_tensor_2,
min=0.,
max=1.,
name="output")
sess = session.Session()
input_arrays_map = [("inputA", [1, 16, 16, 3]),
("inputB", [1, 16, 16, 3])]
output_arrays = ["output"]
with self.assertRaises(ValueError) as error:
convert.toco_convert_graph_def(sess.graph_def,
input_arrays_map,
output_arrays,
enable_mlir_converter=False,
inference_type=dtypes.uint8)
self.assertEqual(
"The `quantized_input_stats` flag must be defined when either "
"`inference_type` flag or `inference_input_type` flag is set to "
"tf.int8 or tf.uint8.", str(error.exception))
def testGraphDefBasic(self):
in_tensor = array_ops.placeholder(
shape=[1, 16, 16, 3], dtype=dtypes.float32, name="input")
_ = in_tensor + in_tensor
sess = session.Session()
tflite_model = convert.toco_convert_graph_def(
sess.graph_def, [("input", [1, 16, 16, 3])], ["add"],
inference_type=lite_constants.FLOAT)
self.assertTrue(tflite_model)
# Check values from converted model.
interpreter = Interpreter(model_content=tflite_model)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
self.assertEqual(1, len(input_details))
self.assertEqual("input", input_details[0]["name"])
self.assertEqual(np.float32, input_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == input_details[0]["shape"]).all())
self.assertEqual((0., 0.), input_details[0]["quantization"])
output_details = interpreter.get_output_details()
self.assertEqual(1, len(output_details))
self.assertEqual("add", output_details[0]["name"])
self.assertEqual(np.float32, output_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == output_details[0]["shape"]).all())
self.assertEqual((0., 0.), output_details[0]["quantization"])
def testGraphDefBasic(self):
in_tensor = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="input")
_ = in_tensor + in_tensor
sess = session.Session()
tflite_model = convert.toco_convert_graph_def(
sess.graph_def, [("input", [1, 16, 16, 3])], ["add"],
inference_type=lite_constants.FLOAT)
self.assertTrue(tflite_model)
# Check values from converted model.
interpreter = Interpreter(model_content=tflite_model)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
self.assertEqual(1, len(input_details))
self.assertEqual("input", input_details[0]["name"])
self.assertEqual(np.float32, input_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == input_details[0]["shape"]).all())
self.assertEqual((0., 0.), input_details[0]["quantization"])
output_details = interpreter.get_output_details()
self.assertEqual(1, len(output_details))
self.assertEqual("add", output_details[0]["name"])
self.assertEqual(np.float32, output_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == output_details[0]["shape"]).all())
self.assertEqual((0., 0.), output_details[0]["quantization"])
def testGraphDefQuantizationInvalid(self):
in_tensor_1 = array_ops.placeholder(
shape=[1, 16, 16, 3], dtype=dtypes.float32, name="inputA")
in_tensor_2 = array_ops.placeholder(
shape=[1, 16, 16, 3], dtype=dtypes.float32, name="inputB")
_ = array_ops.fake_quant_with_min_max_args(
in_tensor_1 + in_tensor_2, min=0., max=1., name="output")
sess = session.Session()
input_arrays_map = [("inputA", [1, 16, 16, 3]), ("inputB", [1, 16, 16, 3])]
output_arrays = ["output"]
with self.assertRaises(ValueError) as error:
convert.toco_convert_graph_def(
sess.graph_def,
input_arrays_map,
output_arrays,
inference_type=lite_constants.QUANTIZED_UINT8)
self.assertEqual(
"std_dev and mean must be defined when inference_input_type is "
"QUANTIZED_UINT8.", str(error.exception))
def testGraphDefQuantization(self):
with ops.Graph().as_default():
in_tensor_1 = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="inputA")
in_tensor_2 = array_ops.placeholder(shape=[1, 16, 16, 3],
dtype=dtypes.float32,
name="inputB")
_ = array_ops.fake_quant_with_min_max_args(in_tensor_1 +
in_tensor_2,
min=0.,
max=1.,
name="output")
sess = session.Session()
input_arrays_map = [("inputA", [1, 16, 16, 3]),
("inputB", [1, 16, 16, 3])]
output_arrays = ["output"]
tflite_model = convert.toco_convert_graph_def(
sess.graph_def,
input_arrays_map,
output_arrays,
enable_mlir_converter=False,
control_output_arrays=None,
inference_type=dtypes.uint8,
quantized_input_stats=[(0., 1.), (0., 1.)])
self.assertTrue(tflite_model)
# Check values from converted model.
interpreter = Interpreter(model_content=tflite_model)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
self.assertEqual(2, len(input_details))
self.assertEqual("inputA", input_details[0]["name"])
self.assertEqual(np.uint8, input_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == input_details[0]["shape"]).all())
self.assertEqual((1., 0.),
input_details[0]["quantization"]) # scale, zero_point
self.assertEqual("inputB", input_details[1]["name"])
self.assertEqual(np.uint8, input_details[1]["dtype"])
self.assertTrue(([1, 16, 16, 3] == input_details[1]["shape"]).all())
self.assertEqual((1., 0.),
input_details[1]["quantization"]) # scale, zero_point
output_details = interpreter.get_output_details()
self.assertEqual(1, len(output_details))
self.assertEqual("output", output_details[0]["name"])
self.assertEqual(np.uint8, output_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == output_details[0]["shape"]).all())
self.assertGreater(output_details[0]["quantization"][0], 0) # scale
def testGraphDefQuantization(self):
in_tensor_1 = array_ops.placeholder(
shape=[1, 16, 16, 3], dtype=dtypes.float32, name="inputA")
in_tensor_2 = array_ops.placeholder(
shape=[1, 16, 16, 3], dtype=dtypes.float32, name="inputB")
_ = array_ops.fake_quant_with_min_max_args(
in_tensor_1 + in_tensor_2, min=0., max=1., name="output")
sess = session.Session()
input_arrays_map = [("inputA", [1, 16, 16, 3]), ("inputB", [1, 16, 16, 3])]
output_arrays = ["output"]
tflite_model = convert.toco_convert_graph_def(
sess.graph_def,
input_arrays_map,
output_arrays,
inference_type=lite_constants.QUANTIZED_UINT8,
quantized_input_stats=[(0., 1.), (0., 1.)])
self.assertTrue(tflite_model)
# Check values from converted model.
interpreter = Interpreter(model_content=tflite_model)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
self.assertEqual(2, len(input_details))
self.assertEqual("inputA", input_details[0]["name"])
self.assertEqual(np.uint8, input_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == input_details[0]["shape"]).all())
self.assertEqual((1., 0.),
input_details[0]["quantization"]) # scale, zero_point
self.assertEqual("inputB", input_details[1]["name"])
self.assertEqual(np.uint8, input_details[1]["dtype"])
self.assertTrue(([1, 16, 16, 3] == input_details[1]["shape"]).all())
self.assertEqual((1., 0.),
input_details[1]["quantization"]) # scale, zero_point
output_details = interpreter.get_output_details()
self.assertEqual(1, len(output_details))
self.assertEqual("output", output_details[0]["name"])
self.assertEqual(np.uint8, output_details[0]["dtype"])
self.assertTrue(([1, 16, 16, 3] == output_details[0]["shape"]).all())
self.assertTrue(output_details[0]["quantization"][0] > 0) # scale
