AttributeError: 'str' object has no attribute 'SerializeToString'

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Traceback (most recent call last):
File “demo.py”, line 87, in
minimum_segment_size=3)
File “/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/tensorrt/python/trt_convert.py”, line 131, in create_inference_graph
input_graph_def_str = input_graph_def.SerializeToString()
AttributeError: ‘str’ object has no attribute ‘SerializeToString’

2

Hello p146103,

Thank you for posting on TensorRT Forum. In order for the community to effectively help you, please describe the issue you are seeing.

What is your platform/GPU? What is the usage scenario? What commands are you executing when you see the error? Any usage/source file you can provide will help us debug too.

regards,
NVIDIA Enterprise Support

3

I have Frozen graph file I want to convert Frozen graph to trt_graph(Tensorrt graph) to speed up my model.
I want to use this trt_graph for object detection Problem

code link:Optimize frozen tensorflow graph using TensorRT · GitHub

Frozen Grpah file:mars-small128.pb - Google Drive

You can suggest me an other code for frozen_graph to tensor_rt grpah

Platform and pakages

Linux:16.04
Gpu:GTX 1080
TensorRt:4.0.1.6
Tensorflow:1.10
Python:3.5
Cuda:9.0

4

Hello,

this seems to be TF specific issue.

def get_frozen_graph(graph_file):
  """Read Frozen Graph file from disk."""
  with tf.gfile.FastGFile(graph_file, "rb") as f:
    graph_def = tf.GraphDef()
[b]    graph_def.ParseFromString(f.read())
[/b]  return graph_def

Can you try TF 1.9 and see if this goes away?