-
Notifications
You must be signed in to change notification settings - Fork 282
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Keras failure in R after using in python with Anaconda #733
Comments
I am working on Windows 10 and R/3.5.0. Apparently Keras has some incompatibility with Python 3.7. The thing is that Tensorflow 1.13.1 is dependent on Python 3.7! I don't know how I can install. I cannot get Tensorflow installed with Python 3.7 and 3.6 on Anacondra. It only works with stand-along installations of Python. Any solution?
|
This may help: rstudio/tensorflow#311 This issue is probably related to reticulate and we already have 2 open issues there eg. rstudio/reticulate#475 and rstudio/reticulate#474 |
This has been solved, see here: rstudio/reticulate#475 |
I have seen the #475 reticulate post, but it does not solve my problem. Re-installing reticulate does not solve the problem. The only solution I have now is to force to downward python 3.7 to python 3.6; then everything seems ok at the moment. |
devtools::install_github("rstudio/reticulate") does not help.
I still got the same error.
FCmodel <- keras_model_sequential()
Error: lexical error: invalid char in json text.
WARNING: The conda.compat modul
(right here) ------^
…On Mon, Apr 8, 2019 at 1:41 PM Daniel Falbel ***@***.***> wrote:
This has been solved, see here: rstudio/reticulate#475
<rstudio/reticulate#475>
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#733 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AURDFn0ckcJc3yqJLFDVkX-7H0Gcob0Wks5ve6lrgaJpZM4chflL>
.
--
Jiancong (Nigel) || 陈见聪
Graduate Student
Environmental Engineering
UC Berkeley
|
Are you on a clean session? Can you share your |
The reason I ran into this is because i updated my Anaconda and python to
3.7. Everything was fine when I was python 3.6.
I just uninstall Anaconda, and R and RStudio. It seemed working now. I
think maybe there’s a compile issue of reticulate in python 3.7
…On Monday, April 8, 2019, Daniel Falbel ***@***.***> wrote:
Are you on a clean session? Can you share your tf_config() and
sessionInfo() please?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#733 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AURDFmwgomTVVYpL36MhmHJ0wSjI1-Cfks5ve-gqgaJpZM4chflL>
.
--
Jiancong (Nigel) || 陈见聪
Graduate Student
Environmental Engineering
UC Berkeley
|
|
Please install the dev version of reticulate:
|
|
This seems like a "permission denied" issue. Are you admin on this machine? |
@dfalbel I am. I just wonder why I get that error with 'conda' json file. |
This error is related to reticulate and has been fixed in the dev version. See here: rstudio/reticulate#474 |
I reinstalled reticulate from GitHub and it worked! So much trouble for a warning :) Cheers. |
Installing reticulate from Github worked. I had to use argument INSTALL_opts = c('--no-lock'). |
Hello,
I have had some weird issues: I mainly used keras on R, but last week I had to do that in python, and then today I realized that my keras installation in RStudio.
I got this error for all keras functions; such as .
model <- keras_model_sequential()
or
rray_reshape(x = xtrain, dim = list(dim(xtrain)[1], 1, dim(xtrain)[2]))
keras Error: lexical error: invalid char in json text. WARNING: The conda.compat module
But when I uninstall anaconda, then I could use my keras.
Does anyone know what may be happening. (I am okay without Anaconda, but it is not that convenient as I do need to use Python occasionally)
The text was updated successfully, but these errors were encountered: