首页 > 解决方案 > TensorFlow toco - 从 *.pb 转换为 *.tflite 错误

问题描述

当尝试使用 TensorFlow 的 toco 转换器将 *.pb 模型转换为 *.tflite(用于在 android 上运行)时,我得到以下信息

2018-05-11 20:38:10.841641: I tensorflow/cc/saved_model/loader.cc:242] Loading SavedModel with tags: { serve }; from: /home/user_name/Downloads/model_directory
2018-05-11 20:38:10.847390: I tensorflow/cc/saved_model/loader.cc:291] SavedModel load for tags { serve }; Status: fail. Took 6042 microseconds.
2018-05-11 20:38:10.848098: F tensorflow/contrib/lite/toco/toco_saved_model.cc:50] **Non-OK-status:** tensorflow::LoadSavedModel(tensorflow::SessionOptions(), tensorflow::RunOptions(), model_path, tags, bundle) **status: Not found:** Could not find meta graph def matching supplied tags: { serve }. To inspect available tag-sets in the SavedModel, please use the SavedModel CLI: `saved_model_cli`Failed to load exported model from /home/user_name/Downloads/model_directory. **Ensure the model contains the required tags 'serve'.**

知道如何解决吗?

机器详情:OS Platform and Distribution - ubuntu x64, python version is 3.5.2, TensorFlow install from - pip, TensorFlow version - cpu version 1.8.0, Bazel version - 0.13.0, CUDA/cuDNN version - no cuda, GPU model和内存 - 没有 gpu,复制的确切命令 - 不需要

标签: androidtensorflowtensorflow-servingtensorflow-lite

解决方案


推荐阅读