tensorflow GPU:资源耗尽错误

时间:2017-06-11 07:32:43

标签: nvidia tensorflow-gpu

我目前正在张量流GPU上运行有关3D医疗图像的CNN。每当我运行代码时,命令提示符上都会显示资源耗尽错误。我已经尝试过小批量1(一次一个病人)运行代码。

我的GPU是NVIDIA GeForce GTX 960.我正在查看我的GPU规格,但我不确定哪个组件限制了内存。是标准内存配置(2GB)吗?

命令提示符返回以下内容:

2017-06-11 16:23:37.095587:W c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ framework \ op_kernel.cc:1152]资源耗尽:分配张量形状时的OOM [3,3,3,8,16] 2017-06-11 16:23:47.096178:W c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:273]分配器(GPU_0_bfc)运行尝试分配365.63MiB的内存不足。

目前的分配摘要如下。 2017-06-11 16:23:47.096349:我c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:643] Bin(256): Total Chunks:0,正在使用的块:0 0B分配给块。 0B客户端请求的块。在bin中使用0B。 0B客户端请求在bin中使用。 ...... 2017-06-11 16:23:47.144036:我c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:696] 3块大小为13824总计40.5KiB 2017-06-11 16:23:47.144745:我c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:696] 3块大小383385600总计1.07GiB 2017-06-11 16:23:47.145486:我c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:696] 1块大小398141184总计379.70MiB 2017-06-11 16:23:47.146771:我c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:700]总和in-使用块:1.44GiB 2017-06-11 16:23:47.146796:我c:\ tf_jenkins \ home \ workspace \ release-win \ device \ gpu \ os \ windows \ tensorflow \ core \ common_runtime \ bfc_allocator.cc:702]

统计:

限制:1548396134

InUse:1548374272

MaxInUse:1548396032

NumAllocs:35

MaxAllocSize:398141184

0 个答案:

没有答案