How to transform tensorflow checkpoint model(.meta .index .data) to frozen model (.pb)

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

Hey guys:

I want to transform this network, https://github.com/jiangsutx/SRN-Deblur, from tensorflow to nvidia tensorRT. It need a ‘.pb’ model file, but the project only giving three model files as follow:

  • deblur.model-52300.data-00000-of-00001
  • deblur.model-52300.index
  • deblur.model-52300.meta

So I want to transform these files to a ‘.pb’ file.

I have tested the ideas given by:

The problem is that both gave ideas are failed because of the get_checkpoint_state() and latest_checkpoint() giving None value.

Is this caused by the missing of checkpoint file?

Are there other ways to implement this?

Any idea is appreciated.

Thanks.

2

Hello,

this is not TensorRT related.

You can check the scripts provided by Google

have a look at this file freeze_graph.py, you can generate .pb file directly from a checkpoint file and a meta graph.
While executing freeze_graph.py, you can specify the input arguments like this:

--input_meta_graph=/path/to/your/meta/graph/model-1.meta  
 --input_checkpoint=/path/to/your/checkpoint/file/model-1  
 --output_graph=/tmp/frozen_graph.pb  
 --output_node_names="Openpose/concat_stage7"  
 --input_binary=True
3

Thanks for your help.

4

hi, in my case I don’t have the .meta file. How can I generate the frozen model ?
output directory of the training looks like this-
training-
-train
-events.out.tfevents.1630519964.DESKTOP-M8813GH.8388.1503.v2
-checkpoint
-ckpt-20.data-00000-of-00001
-ckpt-20.index

5

Hi,
We recommend you to check the below samples links in case of tf-trt integration issues.

If issue persist, We recommend you to reach out to Tensorflow forum.
Thanks!