TensorFlow GPU:ran out of memory trying to allocate 1.14GiB

参考：《深度学习图像识别技术--基于TensorFlow Object Detection API 和 OpenVINO》

        在使用比较低阶的GPU(例如笔记本电脑，GeForce MX150)，训练TensorFlow 模型是，经常会遇到一个错误： Allocator (GPU_0_bfc) ran out of memory trying to allocate 1.14GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available. 如下图所示：

       这时，请减少batch_size

若不愿意减少batch_size,可以切换为CPU，若内存比较大的话。

参考文献：

《How can I solve 'ran out of gpu memory' in TensorFlow》

《Tensorflow：使用 GPU》