You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Mar 3, 2024. It is now read-only.
I am trying to learn BERT with R. I loaded "reticulate" and use reticulate::py_install to install keras-bert. Then a line of code k_bert = import('keras_bert') was run successfully. However, when trying with
model %>% compile(
k_bert$AdamWarmup(decay_steps=decay_steps,
warmup_steps=warmup_steps, learning_rate=learning_rate),
loss = 'binary_crossentropy',
metrics = 'accuracy'
)
and throws errors:
Error in py_call_impl(callable, call_args$unnamed, call_args$named) :
AttributeError: 'AdamWarmup' object has no attribute '_set_hyper'
Traceback detail takes me to the warmup_v2.py script.
I am using Mac OS 12.6.3
R 4.3.1
keras_bert 0.89.0
Thank you in advance.
A clear and concise description of what the bug is.
Version Info
I'm using the latest version
Minimal Codes To Reproduce
importkeras_bertpass
The text was updated successfully, but these errors were encountered:
I am trying to learn BERT with R. I loaded "reticulate" and use reticulate::py_install to install keras-bert. Then a line of code
k_bert = import('keras_bert')
was run successfully. However, when trying withwhere decay_steps = 29, warmup_steps = 2, learning_rate = 1 e-04
The example code calls warmup_v2.py which is located in the following directory:
Library/r-miniconda/envs/r-reticulate/lib/python3.9/site-packages/keras_bert/optimizers/warmup_v2.py
and throws errors:
Error in py_call_impl(callable, call_args$unnamed, call_args$named) :
AttributeError: 'AdamWarmup' object has no attribute '_set_hyper'
Traceback detail takes me to the warmup_v2.py script.
I am using Mac OS 12.6.3
R 4.3.1
keras_bert 0.89.0
Thank you in advance.
A clear and concise description of what the bug is.
Version Info
Minimal Codes To Reproduce
The text was updated successfully, but these errors were encountered: