关于在Windows单卡训练问题

[STRIVESS]

Lane-Segmentation-Solution-For-BaiduAI-Autonomous-Driving-Competition/train.py

Line 63 in 7b13d85

batch_size = 2

1. 是不是在Windows单卡训练的时候，把batch_size = 2改为batch_size = 1？
2. 请问一下，你当时运行的环境是在Ubuntu系统下运行的吗？因为你的环境是paddlepaddle-gpu 1.3.0.post97
   ，就是paddle 1.3.0版本+cuda9.0+cudnn v7+双卡训练（我了解到paddle安装教程中，只有linux系统才支持双卡训练，Windows只支持单卡训练）

3.我的软件环境为 cuda_9.0.176_windows、cudnn-9.0-windows7-x64-v7.6.5.32，依 赖 Python 3. 6、硬件环境为四核 i5-6300HQ_@_2.30GHz、8G RAM、NVIDIA GeForce 950 M( 2 GB )。我已经配置好train.py文件，训练时候就出现了以下问题：

W0317 20:51:26.288599 74176 device_context.cc:362] Please NOTE: device: 0, GPU Compute Capability: 5.0, Driver API Version: 9.0, Runtime API Version: 9.0
W0317 20:51:26.379599 74176 device_context.cc:372] device: 0, cuDNN Version: 7.6.
W0317 20:51:57.438599 74176 operator.cc:206] batch_norm raises an exception struct paddle::memory::allocation::BadAlloc, ResourceExhaustedError:

RuntimeError: ResourceExhaustedError:
Out of memory error on GPU 0. Cannot allocate 24.000244MB memory on GPU 0, available memory is only 5.976563MB.

Please check whether there is any other process using GPU 0.

1. If yes, please stop them, or start PaddlePaddle on another GPU.
2. If no, please decrease the batch size of your model.

(at D:\v2.0.0\paddle\paddle\fluid\memory\allocation\cuda_allocator.cc:69)

请问怎样解决？感谢你百忙之中的回复，感激不尽