To call `multi_gpu_model` with `gpus=2`, we expect the following devices to be available: [‘/cpu:0‘,

解决Keras多GPU配置问题

最新推荐文章于 2023-12-28 11:03:31 发布

原创最新推荐文章于 2023-12-28 11:03:31 发布 · 4.1k 阅读

2 ·

CC 4.0 BY-SA版权

AWS深度学习部署专栏收录该内容

11 篇文章

订阅专栏

本文探讨了在使用Keras框架尝试配置多GPU时遇到的常见错误：ValueError。详细解析了错误信息，提供了检查GPU可用性的Python代码，并指导读者如何通过更新GPU驱动来解决问题，确保所有GPU都能被正确识别。

如何使用一个GPU

To call `multi_gpu_model` with `gpus=2`, we expect the following devices to be available: ['/cpu:0', '/gpu:0', '/gpu:1']. However this machine only has: ['/cpu:0']. Try reducing `gpus

ValueError: To call `multi_gpu_model` with `gpus=2`, we expect the following devices to be available: ['/cpu:0', '/gpu:0', '/gpu:1']. However this machine only has: ['/cpu:0']. Try reducing `gpus`.

It looks like Keras only sees one of the GPUs.

Make sure that all 4 GPUs are accessible, you can use device_lib with TensorFlow.

from tensorflow.python.client import device_lib

def get_available_gpus():
    local_device_protos = device_lib.list_local_devices()
    return [x.name for x in local_device_protos if x.device_type == 'GPU']

You might need to manually install or update GPU drivers on your instance. Consult here.

https://stackoverflow.com/questions/52950449/valueerror-when-using-multi-gpu-model-in-keras