如何解决运行时错误：在进行训练后量化（来自saved_model 的完全量化的tflite 模型）时张量Cast 的空最小值/最大值？

Question

I try to create fully quantized tflite model to be able to run it on coral. I downloaded SSD MobileNet V2 FPNLite 640x640 from https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md

I installed in virtual environment tf-nightly-2.5.0.dev20201123 tf-nightly-models and tensorflow/object_detection_0.1

I run this code to do post training quantization

import tensorflow as tf
import cv2
import numpy as np
converter = tf.lite.TFLiteConverter.from_saved_model('./0-ssd_mobilenet_v2_fpnlite_640x640_coco17_tpu-8/saved_model/',signature_keys=['serving_default']) # path to the SavedModel directory

VIDEO_PATH = '/home/andrej/Videos/outvideo3.h264'
def rep_data_gen():
    REP_DATA_SIZE = 10#00
    a = []
    video = cv2.VideoCapture(VIDEO_PATH)
    i=0
    while(video.isOpened()): 
        ret, img = video.read()
        i=i+1
        if not ret or i > REP_DATA_SIZE:
          print('Reached the end of the video!')
          break
        img = cv2.resize(img, (640, 640))#todo parametrize based on network size
        img = img.astype(np.uint8)
        #img = (img /127.5) -1 #
        #img = img.astype(np.float32)#causing types mismatch error
        a.append(img)
    a = np.array(a)
    print(a.shape) # a is np array of 160 3D images
    for i in tf.data.Dataset.from_tensor_slices(a).batch(1).take(REP_DATA_SIZE):
        yield [i]

#tf2 models
converter.allow_custom_ops=True
converter.optimizations = [tf.lite.Optimize.DEFAULT]
converter.representative_dataset = rep_data_gen
converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS_INT8, tf.lite.OpsSet.SELECT_TF_OPS]
#converter.quantized_input_stats = {'inputs': (0, 255)} #does not help
converter.inference_input_type = tf.uint8  # or tf.uint8
converter.inference_output_type = tf.uint8  # or tf.uint8
quantized_model = converter.convert()

# Save the model.
with open('quantized_model.tflite', 'wb') as f:
  f.write(quantized_model)

I got

RuntimeError: Max and min for dynamic tensors should be recorded during calibration: Failed for tensor Cast
Empty min/max for tensor Cast

Answer 1

I trained the same model, SSD MobileNet V2 FPNLite 640x640 , using the script model_main_tf2.py and then exported the checkpoint to saved_model using the script exporter_main_v2.py . When trying to convert to ".tflite" for use on Edge TPU I was having the same problem.

The solution for me was to export the trained model using the script export_tflite_graph_tf2.py instead of exporter_main_v2.py to generate the saved_model.pb . Then the conversion occurred well.

Maybe try to generate a saved_model using export_tflite_graph_tf2.py .