Ubuntu16.04 Tensorflow的pb模型转tflite模型

最新推荐文章于 2022-11-05 13:08:11 发布

QQ515311445

最新推荐文章于 2022-11-05 13:08:11 发布

阅读量267

点赞数

分类专栏： Linux 图像算法人工智能

本文链接：https://blog.csdn.net/llxxhm/article/details/85329691

版权

Linux 同时被 3 个专栏收录

83 篇文章 0 订阅

订阅专栏

图像算法

11 篇文章 0 订阅

订阅专栏

人工智能

10 篇文章 0 订阅

订阅专栏

1. 下载tensorflow源码

git clone https://github.com/tensorflow/tensorflow

2. 安装bazel

apt-get install openjdk-8-jdk
echo "deb [arch=amd64] http://storage.googleapis.com/bazel-apt stable jdk1.8" | sudo tee /etc/apt/sources.list.d/bazel.list
curl https://bazel.build/bazel-release.pub.gpg | sudo apt-key add -
apt-get update
apt-get install bazel
apt-get upgrade bazel

3. 编译summarize_graph工具

cd tensorflow
bazel build tensorflow/tools/graph_transforms:summarize_graph

4. 查看pb模型的input_arrarys, input_shapes, output_arrarys等信息

bazel-bin/tensorflow/tools/graph_transforms/summarize_graph --in_graph=/tmp/output_graph.pb

5. pb转tflite

tflite_convert --graph_def_file=/tmp/output_graph.pb \
                --output_file=/tmp/output_graph.tflite \
                --input_format=TENSORFLOW_GRAPHEDF \
                --output_format=TFLITE \
                --input_shape=1,299,299,3 \
                --input_array=Placeholder \
                --output_array=final_result \
                --inference_type=FLOAT \
                --input_data_type=FLOAT

或者使用以下python转换

import tensorflow as tf
path="output_graph.pb"
inputs=["Placeholder"]
classes=["final_result"]
converter = tf.contrib.lite.TocoConverter.from_frozen_graph(path, inputs, classes)
tflite_model=converter.convert()
open("output_graph.tflite","wb").write(tflite_model)

6. Python测试

bazel build tensorflow/lite/examples/python:label_image
bazel-bin/tensorflow/lite/examples/python/label_image \
                --image /home/test.jpg \
                --model_file /tmp/output_graph.tflite \
                --label_file /tmp/output_labels.txt

7. C测试(注:只支持bmp)

bazel build tensorflow/lite/examples/label_image:label_image
bazel-bin/tensorflow/lite/examples/label_image/label_image \
                --image=/home/test.bmp \
                --tflite_model=/home/output_graph.tflite \
                --labels=/home/output_labels.txt