Tensorflow 对象检测 API 变量在检查点中不可用

Question

I am trying to retrain an object detection model from Tensorflow's model zoo , but I am getting a looong list of variables that were not found in the checkpoint (see some of the warnings in the list below).

I am using:

• Tensoflow version: 1.15.2
• Windows 10
• Python 3.6.10

The config file that I am using is the one that comes in the zip file with the pre-trained model and the only thing that I have changed is number of detected classes and the paths.

I tried multiple models, but I get the same problem every time.

The transfer learning works, but I guess that I'm not using a significant number of pre-trained weights. What is causing that?

Full log can be found here.

My config file can be found here

Warnings:

W0215 22:04:03.197386  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/BatchNorm/beta/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.197386  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/BatchNorm/beta/RMSProp] is not available in checkpoint
W0215 22:04:03.197386  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/BatchNorm/beta/RMSProp_1] is not available in checkpoint
W0215 22:04:03.197386  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/BatchNorm/gamma/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.197386  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/BatchNorm/gamma/RMSProp] is not available in checkpoint
W0215 22:04:03.197386  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/BatchNorm/gamma/RMSProp_1] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/weights/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/weights/RMSProp] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_0/weights/RMSProp_1] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/BatchNorm/beta/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/BatchNorm/beta/RMSProp] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/BatchNorm/beta/RMSProp_1] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/BatchNorm/gamma/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/BatchNorm/gamma/RMSProp] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/BatchNorm/gamma/RMSProp_1] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/depthwise_weights/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/depthwise_weights/RMSProp] is not available in checkpoint
W0215 22:04:03.198379  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_depthwise/depthwise_weights/RMSProp_1] is not available in checkpoint
W0215 22:04:03.199381  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_pointwise/BatchNorm/beta/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.199381  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_pointwise/BatchNorm/beta/RMSProp] is not available in checkpoint
W0215 22:04:03.199381  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_pointwise/BatchNorm/beta/RMSProp_1] is not available in checkpoint
W0215 22:04:03.199381  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_pointwise/BatchNorm/gamma/ExponentialMovingAverage] is not available in checkpoint
W0215 22:04:03.199381  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_pointwise/BatchNorm/gamma/RMSProp] is not available in checkpoint
W0215 22:04:03.199381  5668 variables_helper.py:157] Variable [FeatureExtractor/MobilenetV1/Conv2d_10_pointwise/BatchNorm/gamma/RMSProp_1] is not available in checkpoint

Answer 1

Not sure if this is still relevant but I also encountered this issue and it looks like this might be a display bug as stated here by one of the contributors to the models repo.

No matter what the base model is ,this bug will happen just because we use pretrained model to train ssd.If we initialize the ssd model randomly, this won't happen.More,it will use more cpu than randomly initialized when training.(About 55%cpu used compared to 45%)