Python tensorflow报错： CUDA_ERROR_OUT_OF_MEMORY

TensorFlow CUDA显存溢出解决方案

最新推荐文章于 2025-05-28 08:14:07 发布

原创

最新推荐文章于 2025-05-28 08:14:07 发布 · 3.6k 阅读

11 ·

CC 4.0 BY-SA版权

文章标签：

#tensorflow #深度学习 #神经网络

tensorflow报错： CUDA_ERROR_OUT_OF_MEMORY

这几天在做卷积神经网络的一个项目，遇到了一个问题CUDA_ERROR_OUT_OF_MEMORY。运行代码时前三四百次都运行正常，之后就一直报这个错误（而且第二次、第三次重新运行程序时，报错会提前），但是程序不停止。今天空闲下来，就看一看这个问题。
详细信息（因为我的报错解决了，所以从网上找了一个相同的报错信息）：

2017-03-28 23:38:19.485396: E tensorflow/stream_executor/cuda/cuda_driver.cc:924] failed to alloc 8589934592 bytes on host: CUDA_ERROR_OUT_OF_MEMORY
2017-03-28 23:38:19.485583: W ./tensorflow/core/common_runtime/gpu/pool_allocator.h:195] could not allocate pinned host memory of size: 8589934592
2017-03-28 23:38:19.485729: E tensorflow/stream_executor/cuda/cuda_driver.cc:924] failed to alloc 7730940928 bytes on host: CUDA_ERROR_OUT_OF_MEMORY
2017-03-28 23:38:19.485823: W ./tensorflow/core/common_runtime/gpu/pool_allocator.h:195] could not allocate pinned host memory of size: 7730940928
2017-03-28 23:38:19.485932: E tensorflow/stream_executor/cuda/cuda_driver.cc:924] failed to alloc 6957846528 bytes on host: CUDA_ERROR_OUT_OF_MEMORY
2017-03-28 23:38:19.486023: W ./tensorflow/core/common_runtime/gpu/pool_allocator.h:195] could not allocate pinned host memory of size: 6957846528
2017-03-28 23:38:19.486165: E tensorflow/stream_executor/cuda/cuda_driver.cc:924] failed to alloc 6262061568 bytes on host: CUDA_ERROR_OUT_OF_MEMORY
2017-03-28 23:38:19.486233: W ./tensorflow/core/common_runtime/gpu/pool_allocator.h:195] could not allocate pinned host memory of size: 6262061568
2017-03-28 23:38:19.486324: E tensorflow/stream_executor/cuda/cuda_driver.cc:924] failed

最低0.47元/天解锁文章