Memory leak and how to train using a gpu ?

lim-anggun / FgSegNet_v2

FgSegNet_v2: "Learning Multi-scale Features for Foreground Segmentation.” by Long Ang LIM and Hacer YALIM KELES

https://arxiv.org/abs/1808.01477

Other

149 stars 43 forks source link

Memory leak and how to train using a gpu ? #17

Open AriouatI opened 4 years ago

AriouatI commented 4 years ago

how can i enable gpu training cause i'm trying to train on colab but it always goes beyond the 12Gb RAM limit is it even normal ? i'm using keras 2.0.6 and tensorflow 1.1.0 as advised but the training process is extremely long and only works with batch_size= 1 otherwise there's a RAM memory problem !

Thanks for your help. @lim-anggun

Akhilesh64 commented 4 years ago

@AriouatI. If you enable gpu runtime the scripts will automatically use gpu. Are you getting this error while extracting masks ???

AriouatI commented 4 years ago

@Akhilesh64 thanks for your comment.

The gpu is enabled but unused (maybe due to tf version used in this code ?) The RAM problem happens either :

during the model.fit() when the batch size > 1 (looks like a memory leak)
during the call to getData.

@lim-anggun

helenypzhang commented 3 years ago

how can i enable gpu training cause i'm trying to train on colab but it always goes beyond the 12Gb RAM limit is it even normal ? i'm using keras 2.0.6 and tensorflow 1.1.0 as advised but the training process is extremely long and only works with batch_size= 1 otherwise there's a RAM memory problem !

Thanks for your help. @lim-anggun

Could you please tell me how to download tensorflow-gpu 1.1.0? Why can't I find it with conda command?