Tiberium

41 ms·

661.
by Tiberium 3y ago
This looks really cool! I really think an idea of a proper UI tool to chain different tools together is nice. https://flowiseai.com/ seems to be the most similar, but it's quite a bit more technical, and doesn't s
662.
by Tiberium 3y ago
Araq has mentioned he'll create a new edition of the book for Nim 2.0, and publish it as an e-book, see https://forum.nim-lang.org/t/10366
663.
by Tiberium 3y ago
All games on Steam have a 2-week after buying or 2 hours after playing refund period, so it's easy to try to see if the game works on Linux or not.
664.
by Tiberium 3y ago
This is not true, I've played it on Linux with Proton myself just fine today. Maybe the poster has a different issue.
665.
by Tiberium 3y ago
This is not actually true for Nim - almost all Nim Discord channels are bridged to different IRC channels, see https://github.com/nim-lang/Nim/wiki#bridged-real-time-chats
666.
Google Colab no longer allows using remote UIs(twitter.com)
19 points by Tiberium 4y ago|5 comments
667.
by Tiberium 4y ago
And a response from Google Colab Product lead - https://twitter.com/thechrisperry/status/1649189902079381505
668.
by Tiberium 4y ago
Isn't it already open-source? https://github.com/MarginaliaSearch/MarginaliaSearch
669.
by Tiberium 4y ago
Very cool project! How much hardware do I need to run this? E.g. can I run it on an old Xeon server with 12 threads and 48GB RAM with HDDs, or that'd be too slow?
670.
by Tiberium 4y ago
Looks cool! Are you planning on adding more customization to be able to influence the AI? See https://bettergpt.chat/ (it's also open source and uses API in the browser). Basically with that frontend you can control th
671.
by Tiberium 4y ago
> and it crashed I hope you understand that it's only because of inability of OpenAI's servers to keep up with demand or some issue in their backend code - language models themselves can't "crash" like normal pro
672.
Memgraph – heap visualisation for multithreaded programs(github.com)
1 points by Tiberium 4y ago|1 comments
673.
by Tiberium 4y ago
"Every allocation is drawn in real time in a little gui window with the relative addresses mapped to 2D hilbert space; where different colors will be used for different threads. The brightness of the allocations fades with time, showin
674.
by Tiberium 4y ago
You're misinterpreting the comment a bit, the poster above is asking for the original Alpaca weights. The code in the repo uses a reproduction of Alpaca with different training code and method of fine-tune (LoRA).
675.
SearchGPT – grounded search engine (Bing Chat alternative)(github.com)
2 points by Tiberium 4y ago|0 comments
676.
Alpaca-LoRA (LLaMA instruct fine-tune) – open-source training code and weights(github.com)
6 points by Tiberium 4y ago|0 comments
677.
by Tiberium 4y ago
But they still didn't publish it into the _public_, like a public download link on their GitHub repo, so it's still a leak.
678.
by Tiberium 4y ago
The source repo seems to be unavailable (private?), so I wanted to ask - is this a language model fine-tuned on your data, or you're using OpenAI's APIs with all data of your own stored in embeddings which are searched for the use
679.
by Tiberium 4y ago
Maybe they're using dedicated OpenAI's Foundry? In the docs they say they're actually using the Turbo model for it, but it seems to be much faster.
680.
by Tiberium 4y ago
Their Sage bot is extremely fast (much faster than gpt-3.5-turbo through OpenAI's API) and seems to be knowledgeable enough about niche topics, very weird.
681.
by Tiberium 4y ago
Such a thing already exists and there were some results - https://open-assistant.io I'm not sure why the authors of Alpaca didn't try to train it on this dataset.
682.
The New OpenAI tokenizer for the Turbo model is much better than the GPT-2/3 one
19 points by Tiberium 4y ago|5 comments
683.
by Tiberium 4y ago
Yes, you are right, every project that distributes LLaMA right now is violating Meta's agreement.
684.
by Tiberium 4y ago
It's actually not.
685.
by Tiberium 4y ago
The original 4chan thread seems to indicate that the leaker verified that his hashes matched with another person who had access to the weights, to make sure that the weights aren't watermarked [0] 0: https://boards.4channel.
686.
by Tiberium 4y ago
It seems that the leak originated from 4chan [1]. Two people in the same thread had access to the weights and verified that their hashes match [2][3] to make sure that the model isn't watermarked. However, the leaker made a mistake of
687.
by Tiberium 4y ago
I think you either have to deal with big prices or actually run the model yourself. There are a lot of GPU renting services that are cheap, like http://coreweave.com/ and https://vast.ai/
688.
by Tiberium 4y ago
> With Windows, there isn't an equivalent ad-free option. You're right, the closest we have is ReactOS which doesn't have enough interest (although understandably so).
689.
by Tiberium 4y ago
Obligatory https://manjarno.snorlax.sh/ For real, advising someone to use Manjaro over Arch is kinda weird.
690.
by Tiberium 4y ago
If anyone's interested, Mullvad has both rented and owned servers - https://mullvad.net/en/help/server-list/ The full list of their servers - https://mullvad.net/en/servers/

More ›