Chunk at 0000001313C1A500 of size 512

最新推荐文章于 2021-04-21 21:30:11 发布

wenrouruwanlian_qy

最新推荐文章于 2021-04-21 21:30:11 发布

阅读量1.4k

点赞数 1

分类专栏： tensorflow

本文链接：https://blog.csdn.net/wenrouruwanlian_qy/article/details/90765325

版权

5 篇文章 0 订阅

订阅专栏

错误现象:

2019-06-04 10:14:12.854238: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200000 of size 256

2019-06-04 10:14:12.856756: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200100 of size 256

2019-06-04 10:14:12.859034: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200200 of size 1536

2019-06-04 10:14:12.861374: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200800 of size 768

2019-06-04 10:14:12.863674: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200B00 of size 512

2019-06-04 10:14:12.866204: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200D00 of size 256

2019-06-04 10:14:12.868513: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B200E00 of size 1280

2019-06-04 10:14:12.871218: I tensorflow/core/common_runtime/bfc_allocator.cc:632] Chunk at 000000130B201300 of size 1536

错误原因：

gpu或者cpu内存不够，超出内存

解决办法：

这是由于训练数据集过大，需要使用 batch

例如batch_size=200 改为 batch_size =50

关注

专栏目录