abakus

Nobody knows cooling satellites better than SpaceX

China?

superficial you say. How about I name these in chinese in my package?

I find the coding standard too long to be taken seriously.

Is hydrogen escaping into space something we should be legitimately concerned about if it became mainstream?

https://en.wikipedia.org/wiki/Diffusion-limited_escape

I hope this or related technology won't be used for cryptomining.

Mixed cargo actually enabled amazing throughputs and efficiency: https://www.reddit.com/r/factorio/comments/m8q8rx/sushi_wago...

How much of the global warming can be explained by this increased sun activity?

Its a good read. The author suggested that maybe its time for change, which makes me curious about what new model does he/she have in mind.

A lot of strategy games are very CPU hungry but not picky about GPU.

I like strategy games and I enjoying modding / editing various aspects of these games. I don't think Cloud gaming will allow me to do these.

What happens to Intel's independent GPU project? With the new nVidia 30xx series it seems that Intel GPU will have a hard time finding its niches.

If you just want to understand the Transformer, here is a clean implementation:

https://github.com/blue-season/pywarm/blob/master/examples/t...

Yes, even here, people are more reluctant to comment on political related news, being afraid of getting downvoted to oblivion.

$350 ? really?

Since we are at the dawn of the space era, I have always wondered if earth mass loss due to massive space exploration will become a problem.

In terms of model expressiveness, I made a functional NN building API for PyTorch (just like keras'), which offers the optimal balance of flexibility and expressiveness: https://github.com/blue-season/pywarm

type annotation is actually a fairly niche functionality. It is bad that most people cannot judge by themselves and just follow the hype.

But you have to factor in the age of this universe. A galaxy scaled intelligence probably haven't even finished a single thought since this universe was born.

The Speed of causality is hand crafted for our scale. It pretty much prohibits conscious beings of different scales to exist.

The more I think about it, the more clear that this universe is a containerized experiment on us. The actual question is, what are they looking for?

My heart desires python then

There is also tensorboardX:

https://github.com/lanpa/tensorboardX

Recently I found that a lot of TF2.0 Keras' functionaly does not support eager execution. This makes Pytorch still significantly easier to prototype with than TF2.0.

If you miss Keras' way of defining NNs, you can use PyWarm: https://github.com/blue-season/pywarm which offers a fully functional NN building API for pytorch.

Very informative article!

I use to work for some company that has a very strict no-merge, only rebase policy, because they want to preserve a straight-line commit history. But I think what they actually need is an algorithm that can draw / print commit history in straight. To me, altering the model for the sake of view kinda of defeated the purposes.

https://github.com/blue-season/pywarm PyWarm is a lightweight, high-level neural network construction API for PyTorch. It enables defining all parts of NNs in the functional way.

It is a nice font, but a bit too wide for my taste.

For a very easy to understand explanation of the transformer and attention mechansim, see:

https://blue-season.github.io/transformer-in-5-minutes/

I am looking for a starting programming language for my kids. I don't really like these lego-like drag-and-drop visual programming "languages" though. This definitely looks like in the right direction. Thanks for sharing!

One way is via comments, like this:

PyWarm, the functional API to build neural networks for PyTorch https://github.com/blue-season/pywarm