gniziemazity / ml-course Goto Github PK

While loading the viewer.html file the web browser crashes (the error message may vary with different browsers). The root cause can be either the web browser V8 engine instance has not enough allocated memory or there are too many open files (.png images).

While the memory error can be mitigated by starting the browser with --disable-extensions --max_old_space_size=8192 the too many open files error may be challenging (see Linux ulimit).

A better and easier approach would be to use the image's loading attribute with the value of lazy, ie:

<img src="..." loading="lazy"/>

To fix this change the createRow function of web/js/display.js file as following:

BEFORE

...
const img=document.createElement('img');
...

AFTER

...
const img=document.createElement('img');
img.setAttribute("loading","lazy");
...

What is the right folder to contribute to the final solution? Would opening a new lesson P10 be reasonable?

Running the npm install on any node directory of the P3-P9 lessons will throw the following error right after generating the dataset successfully:

Error: ENOENT: no such file or directory, open '../common/js_objects/samples.js'

The dataset_generator.js:61 expects the common/js_objects directory to exist. While it exists for the P2 lesson, it doesn't for the other lessons.

The solution would be to create the common/js_objectsdirectory (if not exists) before running the npm install.