minimalist | minimalist mailing list manager | File Utils library

I've been trying to translate some PyTorch code to TensorFlow 2, but the TF2 code is around 10 times slower. I've tried looking at where this might come from, and as far as I can tell it comes from the tape.gradient call (performance was the same with keras' .fit function). I've tried to use different data loaders, ways of declaring the model, installations, etc... and the results have been consistent.

Any explanation / solution as to why this is happening would be much appreciated.

Here is a minimalist version of the TF2 code:

I'd like to use envsubst with nginx docker alpine. Documentation:

Using environment variables in nginx configuration (new in 1.19)

Out-of-the-box, nginx doesn't support environment variables inside most configuration blocks. But this image has a function, which will extract environment variables before nginx starts.

Here is an example using docker-compose.yml:

web: image: nginx volumes:

• ./templates:/etc/nginx/templates ports:
• "8080:80" environment:
• NGINX_HOST=foobar.com
• NGINX_PORT=80

By default, this function reads template files in /etc/nginx/templates/*.template and outputs the result of executing envsubst to /etc/nginx/conf.d.

I have a nginx container service of the form:

I want to draw a stack of almost transparent white rectangles that overlap each other.

Each rectangle has an opacity of 0.01

I have 100 rectangles overlaping, I expect the output result to be the sum of all the opacity. In other term, I expect the result to be white without any opacity.

But it's not the case.

1. Why ?
2. How to get the result I want ?

Here is a minimalist code to illustrate the problem

For an application I had to develop a simple telnet module. (Which I could do quickly with the help of minimalistic telnet) I did create a telnet server and a client and I used PuTTY to test the behavior of the server. PuTYY was configured for 'Telnet negotiation mode: active'.

In the server I did see the expected telnet negotiation sequence (and it was giving correct replies). The last action in the PuTTY negotiation was a strange one.

0xff 0xfe 0x32 or IAC DONT 0x32

The third byte, 0x32, is the option and that is where the problem is (for me). I did look up all Telnet options at IANA and according to them the options range 50-137 is unassigned (0x32 is 62).

What is behind PuTTY option 0x32? I have looked on the internet but can't find anything. I have looked up the PuTTY documentation, same thing. Putting up a request to the PuTTY development is severely discouraged due to capacity problems (not enough hands to type a sensible reply).

It is perfectly sensible not to use this unknown option but it still is nagging me. What is option 0x3e supposed to do?

I have the following minimalistic code that works perfectly fine: a continuous while loop keeps plotting my data and if I press the escape key the plotting stops. Now, if one closes the matplotlib-window a new appears because of the plt.pause command, but now the key_event is not attached anymore. Is there a way to keep the connection of new appearing windows and the key_event?

Code:

The MDN page on colgroup indicates that col width is deprecated, however I have not found an alternative when colspan is involved, and you need to specify the width of a "colspanned" column.

Below is a rather minimalistic snippet to illustrate the issue, for a table with three columns and two rows. The middle column is never explicited with a "td" elements.

With a colgroup, it is possible to specify its width, and then everything is well. Without colgroup, the HTML rendering engine is unable to solve the equation, and column widths render incorrectly.

I'm trying to create a minimalistic command line gui with ncurses on Android. I cross compiled ncurses for ARM64 etc and everything seems to run. However; when starting my test 'hello world' ncurses program in adb shell, I get the following error:

So I've been doing some tests regarding enconding errors that I've seen around ñ charcater using CURL, Qt and sever side PHP. I've finally got a super minimalistic example where the error is in the Qt side ONLY. Maybe some one can help me out.

The Qt Code is as follows:

I have a simple tf.keras model: