mo_tf.exe error

[dlliu0422]

I use below command to convert .pb file to .xml and bin file.
"python mo_tf.py --input_model frozen_inference_graph.pb"
The frozen_inference_graph.pb file is download pretrain model of faster_rcnn_inception_v2_coco.
It show below error message:

WARNING: Logging before flag parsing goes to stderr.
E0918 18:03:22.582442 11896 infer.py:160] Shape [-1 -1 -1 3] is not fully defin
ed for output 0 of "image_tensor". Use --input_shape with positive integers to o
verride model input shapes.
E0918 18:03:22.582442 11896 infer.py:180] Cannot infer shapes or values for node
"image_tensor".
E0918 18:03:22.582442 11896 infer.py:181] Not all output shapes were inferred or
fully defined for node "image_tensor".
For more information please refer to Model Optimizer FAQ (https://docs.openvino
toolkit.org/latest/_docs_MO_DG_prepare_model_Model_Optimizer_FAQ.html), question
#40.
E0918 18:03:22.582442 11896 infer.py:182]
E0918 18:03:22.583442 11896 infer.py:183] It can happen due to bug in custom sha
pe infer function <function Parameter.init.. at 0x000000003B
AEC158>.
E0918 18:03:22.583442 11896 infer.py:184] Or because the node inputs have incorr
ect values/shapes.
E0918 18:03:22.583442 11896 infer.py:185] Or because input shapes are incorrect
(embedded to the model or passed via --input_shape).
E0918 18:03:22.589442 11896 infer.py:194] Run Model Optimizer with --log_level=D
EBUG for more information.
E0918 18:03:22.589442 11896 main.py:307] Exception occurred during running repla
cer "REPLACEMENT_ID" (<class 'extensions.middle.PartialInfer.PartialInfer'>): St
opped shape/value propagation at "image_tensor" node.
For more information please refer to Model Optimizer FAQ (https://docs.openvino

what's wrong ?