mstokholm

> I’m based on OpenAI’s GPT-4 architecture, which is a proprietary large language model. It’s designed to understand and generate human-like text across many topics and tasks. While the underlying model itself isn’t open source, it’s widely used for applications requiring advanced natural language understanding and generation. If you want, I can share more about how it works or about alternatives in the AI landscape!

I also use a similar approach for my open source Python automation project, alkymi: https://github.com/MathiasStokholm/alkymi/blob/develop/CHANG...

In addition to the author's comments, I would suggest also using the format from https://keepachangelog.com/en/1.1.0/ (more sub-headings, e.g. "added").

Lastly, pruning the CHANGELOG upon adding and removing stuff again is a great idea to keep entries meaningful for the reader.

This is mostly a non-issue now - I'm playing Baldur's Gate 3, Dota 2, Warcraft III and everything else on Ubuntu 22.04 now thanks to Steam and Proton. Performance is awesome.

alkymi (https://github.com/MathiasStokholm/alkymi) is such a library (disclaimer: I'm the author). I built alkymi to help me cache complex data pipelines, while still having all the niceties of Python.

Hi HN! My Corona project has finally come far enough for me to show it to the world :)

alkymi is a pure Python (3.5+) library for describing and executing tasks and pipelines with built-in caching and conditional evaluation based on checksums.

I started writing alkymi because I was tired of using Make or similar tools to automate testing/validation of algorithms (data -> algorithm -> report). I've found that although most of my data preparation code, as well as the report generation code, is written in Python, I end up having to wrap all the Python code as CLI scripts to interface with Make. This breaks testing, type checking and generally just makes it really easy to make mistakes (no IDE autocomplete!).

You can view the docs at https://alkymi.readthedocs.io/en/latest/. The docs also contain examples, such as https://alkymi.readthedocs.io/en/latest/examples/mnist.html.

I really hope you find the project interesting, and am looking forward to feedback and (hopefully!) having someone else give it a try.

Well, just because they're producing something useful doesn't mean they can't give back at the same time! Small changes add up.

Keep in mind that these are two slightly different things. Masks might not reduce your personal risk of infection, but they might reduce the risk of you infecting others. In either case, you would see a reduction in the total amount of transmission.

Sam Harris has a great podcast about nuclear launch codes (among many other things) and how the current setup is really not that great: https://samharris.org/podcasts/210-logic-doomsday/

Scrolling past animation worked for me on Firefox

Is this entire talk up on YouTube somewhere?

I would suggest you try out Miniconda (https://docs.conda.io/en/latest/miniconda.html). It comes with just the basics, and let's you install TF with GPU support by simply doing:

conda install -c anaconda tensorflow-gpu