我正在尝试使用tensorflow.
想法:我已经编辑了几张图像,并填写了一些脏话。然后,我用imglabel
标记了图片,以便在测试开始后就可以识别出脏话。
但是测试没有成功。应该创建的所谓“图形”不是按图形顺序创建的。
我已经将自己定位于本教程:https://www.youtube.com/watch?v=rWFg6R5ccOc&t=1832s
请帮助!我不知道如何解决pycharm所显示的这些问题:
2019-06-19 10:23:27.875208:我 tensorflow / core / platform / cpu_feature_guard.cc:141]您的CPU支持 TensorFlow二进制文件未编译使用的指令:AVX2 警告:tensorflow:从 C:\ Users \ Elias \ Documents \ TensorFlow_Tut_2_Classification_Walk-through \ venv \ lib \ site-packages \ tensorflow \ python \ training \ saver.py:1266: checkpoint_exists(来自 tensorflow.python.training.checkpoint_management)已弃用, 将在以后的版本中删除。更新说明:使用 标准文件API来检查带有此前缀的文件。追溯 (最近通话最近):文件 “ C:/用户/ Elias /文档/TensorFlow_Tut_3_Object_Detection_Walk-through/4_export_inference_graph.py”, 第108行,在 tf.app.run()文件“ C:\ Users \ Elias \ Documents \ TensorFlow_Tut_2_Classification_Walk-through \ venv \ lib \ site-packages \ tensorflow \ python \ platform \ app.py”, 正在运行的第125行 _sys.exit(main(argv))文件“ C:/Users/Elias/Documents/TensorFlow_Tut_3_Object_Detection_Walk-through/4_export_inference_graph.py”, 主线56 exporter.export_inference_graph(INPUT_TYPE,trainEvalPipelineConfig,TRAINED_CHECKPOINT_PREFIX_LOC,OUTPUT_DIR, input_shape)文件 “ C:\ TensorFlow \ models \ research \ object_detection \ exporter.py”,行 475,在export_inference_graph中 write_inference_graph = write_inference_graph)文件“ C:\ TensorFlow \ models \ research \ object_detection \ exporter.py”,行 404,在_export_inference_graph中 经过训练的_checkpoint_prefix = checkpoint_to_use)文件“ C:\ TensorFlow \ models \ research \ object_detection \ exporter.py”,行 313,在write_graph_and_checkpoint中 saver.restore(sess,trained_checkpoint_prefix)文件“ C:\ Users \ Elias \ Documents \ TensorFlow_Tut_2_Classification_Walk-through \ venv \ lib \ site-packages \ tensorflow \ python \ training \ saver.py”, 1268行,还原中 + compat.as_text(save_path))ValueError:传递的save_path不是有效的检查点: C:\ Users \ Elias \ Documents \ TensorFlow_Tut_3_Object_Detection_Walk-through / training_data / model.ckpt-500
2019-06-19 10:23:27.875208:我 tensorflow / core / platform / cpu_feature_guard.cc:141]您的CPU支持 TensorFlow二进制文件未编译使用的指令:AVX2 警告:tensorflow:从 C:\ Users \ Elias \ Documents \ TensorFlow_Tut_2_Classification_Walk-through \ venv \ lib \ site-packages \ tensorflow \ python \ training \ saver.py:1266: checkpoint_exists(来自 tensorflow.python.training.checkpoint_management)已弃用, 将在以后的版本中删除。更新说明:使用 标准文件API来检查带有此前缀的文件。追溯 (最近通话最近):文件 “ C:/用户/ Elias /文档/TensorFlow_Tut_3_Object_Detection_Walk-through/4_export_inference_graph.py”, 第108行,在 tf.app.run()文件“ C:\ Users \ Elias \ Documents \ TensorFlow_Tut_2_Classification_Walk-through \ venv \ lib \ site-packages \ tensorflow \ python \ platform \ app.py”, 正在运行的第125行 _sys.exit(main(argv))文件“ C:/Users/Elias/Documents/TensorFlow_Tut_3_Object_Detection_Walk-through/4_export_inference_graph.py”, 主线56 exporter.export_inference_graph(INPUT_TYPE,trainEvalPipelineConfig,TRAINED_CHECKPOINT_PREFIX_LOC,OUTPUT_DIR, input_shape)文件 “ C:\ TensorFlow \ models \ research \ object_detection \ exporter.py”,行 475,在export_inference_graph中 write_inference_graph = write_inference_graph)文件“ C:\ TensorFlow \ models \ research \ object_detection \ exporter.py”,行 404,在_export_inference_graph中 经过训练的_checkpoint_prefix = checkpoint_to_use)文件“ C:\ TensorFlow \ models \ research \ object_detection \ exporter.py”,行 313,在write_graph_and_checkpoint中 saver.restore(sess,trained_checkpoint_prefix)文件“ C:\ Users \ Elias \ Documents \ TensorFlow_Tut_2_Classification_Walk-through \ venv \ lib \ site-packages \ tensorflow \ python \ training \ saver.py”, 1268行,还原中 + compat.as_text(save_path))ValueError:传递的save_path不是有效的检查点: C:\ Users \ Elias \ Documents \ TensorFlow_Tut_3_Object_Detection_Walk-through / training_data / model.ckpt-500