Rendering is not working

[sebastian-gerdes]

Dear RStudio team,

thanks a lot for providing the tutorials. This is a good starting point for me.

I copied the source file into a new quarto document in RStudio. When I run the chunks individually, everything is fine. However, when I hit the "render" button, the chunk in which the layers are set up seems to cause a problem - the process is terminated without reporting a specific cause (see screenshot). I have had the same problem with other tutorials as well.

This is the code that seems to be causing the problem: model <- keras_model_sequential() model %>% layer_flatten(input_shape = c(28, 28)) %>% layer_dense(units = 128, activation = 'relu') %>% layer_dense(units = 10, activation = 'softmax')

When I run knit('basic_image_classification.qmd') from the console, a .txt in markdown is compiled which I can then render interactively in RStudio after having changed the suffix to .md.

Any ideas what could be causing the problem?

Any help is greatly appreciated!

Best greetings, Sebastian

PS: Just a suggestion - it could be helpful to give names to the code chunks - this would make talking about the code easier!

[t-kalinowski]

Naming code chunks is a great idea. Contributions welcome :)

To render the documents at the console you'll have to use quarto, either from the terminal like this:

quarto render tutorials/keras/classification.qmd

or from the R console using the quarto package:

library(quarto)
quarto_render("tutorials/keras/classification.qmd")

[sebastian-gerdes]

Dear Tomasz,

thanks for your answer! The problem was also due to an old version of tensorflow in the environment I was working in I just figured out.

I am kind of new to git - what is the best way to contribute / suggest changes to an individual file? Can I somehow "pull" it in my RStudio environment and send something like a "pull request" with the changes to you? I hope I am not "messing around" too much with the git terminology.

Best greetings, Sebastian

[t-kalinowski]

No worries, the git terminology is certainly confusing at first, but becomes clearer with a little explanation. You might enjoy this getting-started guide that shows you how to open your first PR to an R package: https://usethis.r-lib.org/articles/pr-functions.html