Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
About per_process_gpu_memory_fraction and multi gpu.
See original GitHub issueHi，I have som problems when I train the model in the multi gpu way. Just like that
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: [u’/cpu:0’]. Try reducing gpus.
But I can train the model in one gpu. And the memory of gpu just cost 105MB, no matter how can i modified the batchsize from 8 to 128. Could you give me some help? Thank you very much.
Issue Analytics
- State:
- Created 5 years ago
- Comments:8 (3 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
I’m running 1.7.
It sounds to me like your GPUs are not being found / used at all. The network takes way more than 105mb. Does
nvidia-smilist your GPU?