pjfin123

How efficiently do GPUs heat? I'm assuming most of the input energy is lost to heat?

Didn't know about that. If the point is to just be able to play around with Solidity and smart contracts it doesn't seem like a big deal. The Discord banning sounds sketchy though so maybe not something to use beyond testing.

There's Cheap Eth which is an Ethereum fork with cheaper prices: https://cheapeth.org/ .

Yeah it seems like the real defense of an obvious attack is "This Ethereum proof of stake chain with all of the wallets from right before the attack is the new Bitcoin".

Exciting time to be alive!

It seems like Ethereum 2.0 style proof of stake taking off could really limit a lot of the harms with crypto mining.

I wonder if Bitcoin could ever migrate to Ethereum technology/proof of stake? It seems like a good idea but the Bitcoin community seems very resistant to change (see block size debate).

If you want something to live on the web long after you've lost interest in it static HTML/GitHub pages seems like the way to go. The chess board I built in high school to learn JavaScript is still going strong: https://github.com/PJ-Finlay/JavaScript-Chess-Board

Very cool, thanks or making this!

Interesting, thanks for the detailed write up.

> The translation of the first sentence conveys the meaning but changes the date from February 7 to October 7 and drops the year.

Like discussed earlier in this thread Argos Translate tokenizes the incoming text into what are essentially words and then passes them to the Transformer network. For situations where the risks of translations that are fluid and seem reasonable but are totally wrong is high neural translation might not make sense even though the quality is generally higher. Instead you would probably want to use rule-based translations like what the Apertium project does.

It was partially trained on EU and UN documents. Interesting observation, less emphasis on parliamentary/legal data vs a data set like OpenSubtitles, which has more casual language use, might give a better performance.

You're right there's surprisingly little data available for English-Chinese. I'm not sure why, it seems like there would be a lot of demand for translating between them.

For the en-zh model copying this is a known issue: https://github.com/argosopentech/argos-translate/issues/4

Pivoting through English isn't inherent to Argos Translate, you could train a French-German model or whatever you want I've just been focusing on training models to add new languages. The ideal strategy is to have models that know multiple languages.

Quoting a previous HN comment:

I think cloud translation is still pretty valuable in a lot of cases since the model for one single direction translation is ~100MB. In addition to having more language options without a large download cloud translations let you use more specialized models for example French to Spanish. I just have a model to and from English for each language and any other translations have to "pivot" through English. For cloud translations you can also use one model with multiple input and output languages which gives you better quality translation between languages that don't have as much data available and lets you support direct translation between a large number of languages. Here's a talk where Google explains how they do this for Google Translate: https://youtu.be/nR74lBO5M3s?t=1682. You could do this locally but it would have its own set of challanges for getting the right model for the languages you want to translate.

> Thinking about it a bit more, it's a bit weird that the failure mode of a weak model would be to regurgitate the input unchanged. I'd rather have expected random Chinese gibberish in that case. Doesn't that mean the model has seen at least a few cases where English sentences were left untranslated in the training data?

This was added last week, it's just not live on libretranslate.com yet:

https://github.com/uav4geo/LibreTranslate/issues/33

The training scripts are just an example for English-Spanish, Opus(http://opus.nlpl.eu/) has data for English-Chinese.

And a PyQt native Desktop app.

There's more in depth discussion of this issue here: https://github.com/uav4geo/LibreTranslate/issues/20

In some cases using all lower case can help avoid this risk if capitalization isn't important.

There are different models for each language pair. Currently there are only pre-trained models to and from English and other language pairs "pivot" through English.

ex:

es -> en ->fr

Chinese is the weakest language pair currently, but I'm currently working on improving it: https://github.com/argosopentech/argos-translate/issues/17

OPUS parallel corpus: http://opus.nlpl.eu/

It's really great, they have a large amount of data and organize it to make it easy to access.

The process for translation is to "tokenize" a stream of characters into "tokens" like this:

"Hello Sarah, what time is it?"

<Hello><_Sarah><,><_what><_time><_is><_it><?>

Then the tokens are used for inference with a Transformer net. There are no guarantees that output will be consistent to even small changes in the input. The network, based on the data it was trained on/luck, has slightly different connections for the <_Sara> token than for the <_Sarah> token leading to different output.

Here's a video of some Linux YouTubers reviewing Argos Translate (https://github.com/argosopentech/argos-translate), the underlying translation library, and getting unexpected outputs: https://www.youtube.com/watch?v=geMs9dxl1N8

This video really helped me understand Python packaging of binaries:

https://youtu.be/02aAZ8u3wEQ

Buying and holding low fee index funds would probably be a more sustainable way to stick it to Wall Street.

Interesting explainer I saw by an Australian economics YouTuber (opposed to the new law):

https://youtu.be/zj2r948rUkQ

Oh wow JACK and Ardour are awesome, didn't know about the Amazon connection

How do you do the auto updates?

What does the future of open-source large neural nets look like? My understanding is GPT-3 takes ~600GB of GPU memory to run inference. Does an open source model just allow you a choice of a handful of cloud providers instead of one?

I don't think people realize how awesome Craigslist is for things that aren't good candidates for being sold on eBay because of shipping costs. I got a free aquarium a few months ago. It's great!

A YouTube video showing neural style transfer to paint the Mona Lisa in a wide variety of Artistic styles. Here's a test picture I made of the Taj Mahal in the style of "Starry Night": https://www.reddit.com/r/deepdream/comments/ktsx3e/taj_mahal...

Is there an easy way to make F-Droid install updates automatically? I use both F-Droid and Google Play on my phone but manual updates are a huge usability pain.

With all of the current discussion I thought this was an interesting video showing the mechanics of "deplatforming" on federated networks like Mastodon.

This was the biggest problem with PeerTube. For Mastodon I clicked the big "join" button on the website and filtered by technology > foss and fosstodon.org was the largest option and seems to have a great community. For PeerTube trying to do the same thing I found some small technology instance with a few hundred videos. I chose peervideo.club because it's the 4th largest (so I'd be in good company if it's taken down) and seemed to be the largest English focused instance. From their about, "Peer Video Club is a chill PeerTube instance most likely just for testing. How long we plan to maintain this instance: ¯\(ツ)/¯. How we will pay for this instance: Owners's own funds". Doesn't seem like a pillar of stability.

My understanding is I could backup (no plans), and restore to another instance. To take my followers with me I would have to leave peervideo.club in a more planned way and set my account to redirect to another. No idea how long the redirect would have to stay up for most of my subscribers' instances to get the message. Even with federated you need some degree of trust that you won't be kicked off or have you instance shut down with no warning.

I made PeerTube and Mastadon accounts today out of curiosity and was generally impressed. Mastodon especially is quite usable (though obviously many fewer users than Twitter), the interface is as good as Twitter (better if you count not spamming you to login). I posted about my open source project and other random stuff and got several thought out, kind, non-spam replies with a new account. I imagine this is because of the barrier to entry and that I made my account on a open source focused instance. If Mastodon/federated social networking became popular I imagine there could be "Eternal September" (https://en.wikipedia.org/wiki/Eternal_September) dynamics though.

I wonder if paradoxically federated communication would let you improve the quality of discourse by more proactively banning people. Mastadon works on a federated system where anyone can run they're own "instance" which can then interact with other instances pretty seamlessly. Since people really can leave if they don't like the rules you could be pretty heavy handed with moderation within individual instances to ban obnoxious people or have an invite only system.

PeerTube was a little rougher around the edges, but still very impressive for being pretty new and doing p2p video streaming. There was more lag than YouTube but with decent internet it's fine. The choices for instances aren't great and storage size and bandwith are limited. Storage/bandwith cost money though and that may just be making what's an invisible cost on YouTube visible.

My Mastodon: https://fosstodon.org/@argosopentech

My PeerTube: https://peervideo.club/video-channels/argosopentech/videos