Float 16 量化 Tflite model 不适用于自定义模型？

Question

I have a custom tensorflow model which I converted into tflite using the Float16 quantization as mentioned here .
But the the input details of the tflite model using the tflite interpreter are

[{'name': 'input_1',
  'index': 0,
  'shape': array([  1, 256, 256,   3], dtype=int32),
  'shape_signature': array([ -1, 256, 256,   3], dtype=int32),
  'dtype': numpy.float32,
  'quantization': (0.0, 0),
  'quantization_parameters': {'scales': array([], dtype=float32),
   'zero_points': array([], dtype=int32),
   'quantized_dimension': 0},
  'sparsity_parameters': {}}]

while the output details are

[{'name': 'Identity',
  'index': 636,
  'shape': array([  7,   1, 256, 256,   1], dtype=int32),
  'shape_signature': array([  7,  -1, 256, 256,   1], dtype=int32),
  'dtype': numpy.float32,
  'quantization': (0.0, 0),
  'quantization_parameters': {'scales': array([], dtype=float32),
   'zero_points': array([], dtype=int32),
   'quantized_dimension': 0},
  'sparsity_parameters': {}}]

Is something wrong with the conversion?

I also received this warning while converting the tf model to tflite

WARNING:absl:Found untraced functions such as _defun_call, _defun_call, _defun_call, _defun_call, _defun_call while saving (showing 5 of 63). These functions will not be directly callable after loading.
WARNING:absl:Found untraced functions such as _defun_call, _defun_call, _defun_call, _defun_call, _defun_call while saving (showing 5 of 63). These functions will not be directly callable after loading.

PS I also tried doing this quantization, but received the same input/output details for that tflite model.

Answer 1

Float16 and post-training quantization don not modify the input/output tensors but the intermediate weight tensors only. The above behavior looks like an intended behavior. If you want to fully quantize your model, try the full integer quantization. You can find the details at here .