如何使用Tensorflow对象检测API继续训练对象检测模型? [英] How to continue training an object detection model using Tensorflow Object Detection API?

问题描述

我正在使用 Tensorflow对象检测API使用转移学习训练对象检测模型.具体来说，我正在使用 ssd_mobilenet_v1_fpn_coco从模型动物园中，并使用提供了示例管道，当然已用与我的培训以及评估tfrecords和标签的实际链接替换了占位符.

I'm using Tensorflow Object Detection API to train an object detection model using transfer learning. Specifically, I'm using ssd_mobilenet_v1_fpn_coco from the model zoo, and using the sample pipeline provided, having of course replaced the placeholders with actual links to my training and eval tfrecords and labels.

我能够使用上述管道成功地在约5000张图像(以及相应的边界框)上训练模型(如果愿意的话，我主要是在TPU上使用Google的ML引擎).

I was able able to successfully train a model on my ~5000 images (and corresponding bounding boxes) using the above pipeline (I'm mainly using Google's ML Engine on TPU, if revelant).

现在，我准备了约2000张图像，并希望继续用这些新图像训练模型，而无需从头开始(训练初始模型花了大约6个小时的TPU时间).我该怎么办?

Now, I prepared an additional ~2000 images, and would like continue training my model with those new images, without restarting from scratch (training the initial model took ~6h of TPU time). How can I do that?

推荐答案

您有两个选择，都需要更改新版本 train_input_reader 的 input_path 数据集:

You have two options, in both you need to change the input_path of the train_input_reader of your new dataset:

1. 在训练配置中指定要微调的检查点时，请指定训练模型的检查点

train_config{
    fine_tune_checkpoint: <path_to_your_checkpoint>
    fine_tune_checkpoint_type: "detection"
    load_all_detection_checkpoint_vars: true
}

1. 只需继续使用与先前模型相同的配置(除了 train_input_reader )和相同的 model_dir .这样，API将创建一个图形并检查 model_dir 中是否已经存在一个检查点并适合该图形.如果是这样，它将恢复并继续训练.

1. Simply keep using the same configuration (except the train_input_reader) with the same model_dir of your previous model. That way, the API will create a graph and will check whether a checkpoint already exists in model_dir and fits the graph. If so - it will restore it and continue training it.

fine_tune_checkpoint_type先前被错误地设置为true，而通常应为检测"或分类"，在这种情况下应为检测".感谢克里希(Krish)的注意.

fine_tune_checkpoint_type was previously set as true by mistake, while it should be "detection" or "classification" in general, and "detection" in this specific case. Thanks Krish for noticing.

这篇关于如何使用Tensorflow对象检测API继续训练对象检测模型?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！