search

shekkizh / FCN.tensorflow

Tensorflow implementation of Fully Convolutional Networks for Semantic Segmentation (http://fcn.berkeleyvision.org)
MIT License
1.25k stars 527 forks source link

how to decrease space used for logging? #41

Closed bhralzz closed 7 years ago

bhralzz commented 7 years ago

Dear all, I have some limitations on hard disk space on shared server, Is there any necessity for logging? Or how I can use some smaller space for logging? How I can decrease the size need for logging? Please guide step by step Any comment appreciated. Thanks

xiaoxionglin commented 7 years ago

you can set parameters of tensorflow how many checkpoints you want to keep. By default it's 5. And one checkpoint is around 2 GB.

bhralzz commented 7 years ago

Is it possible to give an example in relation to this case?

xiaoxionglin commented 7 years ago

I am not an expert. Just encountered the similar problem when implementing this code. there is sth in folder "logs" called: model.ckpt-500.data-00000-of-00001 as such. it should be the main model data. by default it saves every 500 iterations, you can change it in the file "FCN.py" near the end. sth like: if iter%500==0

The thing I mentioned in the last post: https://www.tensorflow.org/versions/r0.11/api_docs/python/state_ops/saving_and_restoring_variables

you can modify the parameter "Max_to_keep" to a lower value, maybe 1 or 2

bhralzz commented 7 years ago

Ok thanks