e2e-coref icon indicating copy to clipboard operation
e2e-coref copied to clipboard

what is the exact tensorflow version used in the codes

Open ghost opened this issue 5 years ago • 4 comments

Hi tensorflow has changed a lot and when I run the codes I get error for every single lines of this code, could you specify exactly the version compatible? thanks

ghost avatar Oct 30 '19 15:10 ghost

Hi I also have the same issue. If I install tensorflow according to your requirements.txt, either I get error due to tensorflow version or I get the library linkage issue. Please reply. Thanks

chanchalIITP avatar Nov 25 '19 05:11 chanchalIITP

pip install tensorflow-gpu==1.14

henryhust avatar Dec 06 '19 05:12 henryhust

Dear henry, Thank you for the reply. Somehow I have resolve the version issue. Right now, my train.py is running indefinitely and giving no output. It seems that, the program is stuck somewhere. Any help would be appreciated. The model is stuck at line 53 of train.py file.

  tf_loss, tf_global_step, _ = session.run([model.loss,

model.global_step, model.train_op])

Thanks and regards, Chanchal Suman

On Fri, Dec 6, 2019 at 11:01 AM henry [email protected] wrote:

pip install tensorflow-gpu==1.14

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kentonl/e2e-coref/issues/77?email_source=notifications&email_token=AGRCCYIK6YYNVNMDJPXZKALQXHPUXA5CNFSM4JG3IGPKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGDCE3Q#issuecomment-562438766, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGRCCYKY6EMUFDVK7I54GITQXHPUXANCNFSM4JG3IGPA .

chanchalIITP avatar Dec 06 '19 05:12 chanchalIITP

tensorflow-gpu: 1.14, cuda: 10.0, gcc: 5.5.0 works well.

WinterShiver avatar Jun 25 '21 06:06 WinterShiver