tensorflow问题排查

最新推荐文章于 2022-11-02 21:45:08 发布

dataastron

最新推荐文章于 2022-11-02 21:45:08 发布

阅读量3.5k

点赞数

分类专栏：深度学习

本文链接：https://blog.csdn.net/dataastron/article/details/79439110

版权

深度学习专栏收录该内容

11 篇文章 0 订阅

订阅专栏

今天执行个很小的文本训练，一上来就报了个错。

看了下好像是内存分配的问题。

2018-03-04 19:51:05.588759: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)

2018-03-04 19:51:05.589444: E tensorflow/stream_executor/cuda/cuda_driver.cc:936] failed to allocate 201.25M (211025920 bytes) from device: CUDA_ERROR_OUT_OF_MEMORY
Train on 151592 samples, validate on 7979 samples
Epoch 1/2
2018-03-04 19:51:06.693381: E tensorflow/stream_executor/cuda/cuda_blas.cc:366] failed to create cublas handle: CUBLAS_STATUS_NOT_INITIALIZED
2018-03-04 19:51:06.694111: E tensorflow/stream_executor/cuda/cuda_blas.cc:366] failed to create cublas handle: CUBLAS_STATUS_NOT_INITIALIZED
2018-03-04 19:51:06.694804: E tensorflow/stream_executor/cuda/cuda_blas.cc:366] failed to create cublas handle: CUBLAS_STATUS_NOT_INITIALIZED