Tensorflow 2.0 GPU不显示输出（可能崩溃）

Question

我有一个Windows 10系统，带有GeForce 930 MX NVIDIA图形卡、CUDA 10.0、CudNN 7.6.3和tensorflow 2.0-gpu。我正在尝试使用以下链接进行对象检测：https://github.com/tensorflow/models/tree/master/research/object_detection 虽然崩溃且未完成，但我的输出没有给出错误。代码如下：

import numpy as np
import os
import six.moves.urllib as urllib
import sys
import tarfile
import tensorflow as tf
import zipfile
import cv2

from collections import defaultdict
from io import StringIO
from matplotlib import pyplot as plt
from PIL import Image
from IPython.display import display
from object_detection.utils import ops as utils_ops
from object_detection.utils import label_map_util
from object_detection.utils import visualization_utils as vis_util
#tf.debugging.set_log_device_placement(True)
physical_devices = tf.compat.v1.config.experimental.list_physical_devices('GPU') 

tf.compat.v1.config.experimental.set_memory_growth(physical_devices[0], True)

utils_ops.tf = tf.compat.v1
print(utils_ops.tf.config.experimental.list_physical_devices('GPU'))

# Patch the location of gfile
tf.gfile = tf.io.gfile

def load_model(model_name):
  base_url = 'http://download.tensorflow.org/models/object_detection/'
  model_file = model_name + '.tar.gz'
  model_dir = tf.keras.utils.get_file(
    fname=model_name, 
    origin=base_url + model_file,
    untar=True)

  model_dir = pathlib.Path(model_dir)/"saved_model"

  model = tf.saved_model.load(str(model_dir))
  model = model.signatures['serving_default']

  return model




# List of the strings that is used to add correct label for each box.
PATH_TO_LABELS = 'data/mscoco_label_map.pbtxt'
category_index = label_map_util.create_category_index_from_labelmap(PATH_TO_LABELS, use_display_name=True)



import pathlib
PATH_TO_TEST_IMAGES_DIR = pathlib.Path('test_images')
TEST_IMAGE_PATHS = sorted(list(PATH_TO_TEST_IMAGES_DIR.glob("*.jpg")))
model_name = 'ssd_mobilenet_v1_coco_2017_11_17'
detection_model = load_model(model_name)

def run_inference_for_single_image(model, image):
  image = np.asarray(image)
  # The input needs to be a tensor, convert it using `tf.convert_to_tensor`.
  input_tensor = tf.convert_to_tensor(image)
  # The model expects a batch of images, so add an axis with `tf.newaxis`.
  input_tensor = input_tensor[tf.newaxis,...]

  # Run inference
  output_dict = model(input_tensor)

  # All outputs are batches tensors.
  # Convert to numpy arrays, and take index [0] to remove the batch dimension.
  # We're only interested in the first num_detections.
  num_detections = int(output_dict.pop('num_detections'))
  output_dict = {key:value[0, :num_detections].numpy() 
                 for key,value in output_dict.items()}
  output_dict['num_detections'] = num_detections

  # detection_classes should be ints.
  output_dict['detection_classes'] = output_dict['detection_classes'].astype(np.int64)

  # Handle models with masks:
  if 'detection_masks' in output_dict:
    # Reframe the the bbox mask to the image size.
    detection_masks_reframed = utils_ops.reframe_box_masks_to_image_masks(
              output_dict['detection_masks'], output_dict['detection_boxes'],
               image.shape[0], image.shape[1])      
    detection_masks_reframed = tf.cast(detection_masks_reframed > 0.5,
                                       tf.uint8)
    output_dict['detection_masks_reframed'] = detection_masks_reframed.numpy()

  return output_dict

def show_inference(model, image_path):
  # the array based representation of the image will be used later in order to prepare the
  # result image with boxes and labels on it.
  image_np = np.array(Image.open(image_path))
  # Actual detection.
  output_dict = run_inference_for_single_image(model, image_np)
  # Visualization of the results of a detection.
  vis_util.visualize_boxes_and_labels_on_image_array(
      image_np,
      output_dict['detection_boxes'],
      output_dict['detection_classes'],
      output_dict['detection_scores'],
      category_index,
      instance_masks=output_dict.get('detection_masks_reframed', None),
      use_normalized_coordinates=True,
      line_thickness=8)

  cv2.imshow(Image.fromarray(image_np))
  print("OK!")

for image_path in TEST_IMAGE_PATHS:
    show_inference(detection_model, image_path)

以下是输出：

(tfflask) C:\Users\DELL\Downloads\models-master\research\object_detection>python test_object.py
2020-06-04 19:15:31.227206: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_100.dll
2020-06-04 19:15:33.830860: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library nvcuda.dll
2020-06-04 19:15:34.246286: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties:
name: GeForce 930MX major: 5 minor: 0 memoryClockRate(GHz): 1.0195
pciBusID: 0000:03:00.0
2020-06-04 19:15:34.251302: I tensorflow/stream_executor/platform/default/dlopen_checker_stub.cc:25] GPU libraries are statically linked, skip dlopen check.
2020-06-04 19:15:34.258134: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1746] Adding visible gpu devices: 0
[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
2020-06-04 19:15:35.890465: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2020-06-04 19:15:35.966988: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties:
name: GeForce 930MX major: 5 minor: 0 memoryClockRate(GHz): 1.0195
pciBusID: 0000:03:00.0
2020-06-04 19:15:35.980010: I tensorflow/stream_executor/platform/default/dlopen_checker_stub.cc:25] GPU libraries are statically linked, skip dlopen check.
2020-06-04 19:15:35.991055: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1746] Adding visible gpu devices: 0
2020-06-04 19:15:37.099838: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1159] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-06-04 19:15:37.102595: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1165]      0
2020-06-04 19:15:37.104199: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1178] 0:   N
2020-06-04 19:15:37.109167: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1304] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 1384 MB memory) -> physical GPU (device: 0, name: GeForce 930MX, pci bus id: 0000:03:00.0, compute capability: 5.0)
2020-06-04 19:15:43.916571: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudnn64_7.dll
2020-06-04 19:15:44.713084: W tensorflow/stream_executor/cuda/redzone_allocator.cc:312] Internal: Invoking ptxas not supported on Windows
Relying on driver to perform ptx compilation. This message will be only logged once.
2020-06-04 19:15:45.661495: W tensorflow/core/common_runtime/bfc_allocator.cc:239] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.06GiB with freed_by_count=0. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.
2020-06-04 19:15:45.930582: W tensorflow/core/common_runtime/bfc_allocator.cc:239] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.09GiB with freed_by_count=0. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.
2020-06-04 19:15:46.118122: W tensorflow/core/common_runtime/bfc_allocator.cc:239] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.15GiB with freed_by_count=0. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.
2020-06-04 19:15:46.517654: W tensorflow/core/common_runtime/bfc_allocator.cc:239] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.27GiB with freed_by_count=0. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.

我似乎无法理解它崩溃的原因，我尝试使用CuDnn 7.4.1和7.4.2，但它给出了一个类似以下内容的错误： CuDNN failed to initialize 7.6.3中似乎没有给出该错误，但它在没有输出的情况下崩溃。请帮忙