无法通过对象检测API运行以较小的分辨率再次重新冻结的图形

Question

我从模型动物园下载了faster_rcnn_inception_v2。其输入分辨率为1024x600。我尝试通过将配置从-

更改为较小的分辨率来“重新冻结”它

keep_aspect_ratio_resizer {
  min_dimension: 600
  max_dimension: 1024
}

到

keep_aspect_ratio_resizer {
  min_dimension: 360
  max_dimension: 640
}

并运行以下命令行-

python object_detection/export_inference_graph.py \
--pipeline_config_path /hdd/motorola/motorola_data/pre_trained/faster_rcnn_inception_v2_coco_2018_01_28/pipeline_small.config \
--trained_checkpoint_prefix /hdd/motorola/motorola_data/pre_trained/faster_rcnn_inception_v2_coco_2018_01_28/model.ckpt \
--output_directory /hdd/motorola/motorola_data/pre_trained/faster_rcnn_inception_v2_coco_2018_01_28_small_reso/

尽管当我在某些输入上运行图形时，出现以下错误-

ValueError: NodeDef mentions attr 'explicit_paddings' not in Op<name=Conv2D; signature=input:T, filter:T -> output:T; attr=T:type,allowed=[DT_HALF, DT_BFLOAT16, DT_FLOAT, DT_DOUBLE]; attr=strides:list(int); attr=use_cudnn_on_gpu:bool,default=true; attr=padding:string,allowed=["SAME", "VALID"]; attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW"]; attr=dilations:list(int),default=[1, 1, 1, 1]>; NodeDef: {{node FirstStageFeatureExtractor/InceptionV2/InceptionV2/Conv2d_1a_7x7/separable_conv2d}} = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], explicit_paddings=[], padding="VALID", strides=[1, 1, 1, 1], use_cudnn_on_gpu=true](FirstStageFeatureExtractor/InceptionV2/InceptionV2/Conv2d_1a_7x7/separable_conv2d/depthwise, FirstStageFeatureExtractor/InceptionV2/Conv2d_1a_7x7/pointwise_weights/read). (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).

这是调整图的输入分辨率大小而不用期望的分辨率重新训练它的正确方法吗？