问题1:Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA

 解决:代码中添加

import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'

问题2:Tensorflow GPU显存占满,而Util为0

解决:代码中添加

import os
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = "8" //这里指定GPU8

代码中同时指定了GPU,避免了其他GPU的显存占用。

参考文章:https://blog.csdn.net/qq_24306353/article/details/79952266

GitHub 加速计划 / te / tensorflow
21
4
下载
一个面向所有人的开源机器学习框架
最近提交(Master分支:1 个月前 )
4f64a3d5 Instead, check for this case in `ResolveUsers` and `ResolveOperand`, by querying whether the `fused_expression_root` is part of the `HloFusionAdaptor`. This prevents us from stepping into nested fusions. PiperOrigin-RevId: 724311958 1 个月前
aa7e952e Fix a bug in handling negative strides, and add a test case that exposes it. We can have negative strides that are not just -1, e.g. with a combining reshape. PiperOrigin-RevId: 724293790 1 个月前
Logo

旨在为数千万中国开发者提供一个无缝且高效的云端环境,以支持学习、使用和贡献开源项目。

更多推荐