python - 断言错误：未从模型格式中正确检测到框架

Question

我正在尝试英特尔低精度优化工具，我正在关注这个 github（https://github.com/intel/lpot/tree/master/examples/tensorflow/object_detection）。当我运行如下量化命令时

bash run_tuning.sh --config=ssd_mobilenet_v1.yaml --input_model=ssd_mobilenet_v1_coco_2018_01_28/frozen_inference_graph.pb --output_model=./tensorflow-ssd_mobilenet_v1-tune.pb

我收到以下错误。

Traceback (most recent call last):
  File "main.py", line 60, in <module>
    evaluate_opt_graph.run()
  File "main.py", line 48, in run
    quantizer.model = self.args.input_graph
  File "/home/uxxxxx/.conda/envs/lpot/lib/python3.7/site-packages/lpot/experimental/component.py", line 334, in model
    self._model = Model(user_model)
  File "/home/uxxxxx/.conda/envs/lpot/lib/python3.7/site-packages/lpot/experimental/common/model.py", line 43, in __new__
    assert False, 'Framework is not detected correctly from model format.'
AssertionError: Framework is not detected correctly from model format.

请帮忙！

Answer 1

这是因为您的输入模型路径需要一些修正。使用 ./ssd 代替 ssd。它对我有用。问候