Saichethan M. Reddy
Saichethan M. Reddy
@amansrivastava17 thanks that solved my current error, but I'm interested in XLNET so is this issue solved (https://github.com/amansrivastava17/embedding-as-service/issues/45) I ran en.encode(['my name is aman']) two times and got different embedding......
@maziyarpanahi is this solved??
@maziyarpanahi so if I wanted to get XLNet embeddings now what should I do? Wait till they merge or is there any other way
Is this issue still open. Can I work on it?
@kyleclo thanks for your response Then how can I get sentence embeddings in TF??
I did it 5 times, sure scores changed a little but not close to reported scores. where am I doing wrong?
25 On Wed, Sep 4, 2019, 9:05 AM thecharm wrote: > how many epochs do you set? > > — > You are receiving this because you authored the thread....
@Amling2017 no even after parameter tuning we were able to get at max 67.5 which is slightly better than ma hovy at al.
@softhuafei change the path ../data/* to your current directory path if it's in the home directory simply remove ".."
@softhuafei probably that's bcoz of your Keras version try `pip install keras==1.1.2` if that doesn't work check Keras Contrib.