Hey HN! Author of the site here. I tried a few tricks to keep the text-generation part of the site up, but even leaning hard on Huggingface's API and bumping time-outs up, it looks like the site is struggling a bit. I'm going to see if there's anything I can do to keep the text-generation part available, but in the meantime, the pre-generated set should stay pretty stable. Not sure if there's much else I can do without burning a hole in my cloud bills — sorry for the troubles!<p>I've put up a more detailed description of how this works on the GitHub - <a href="https://github.com/thesephist/modelexicon" rel="nofollow">https://github.com/thesephist/modelexicon</a><p>PS - if anyone at Huggingface is reading this and wants to help out with keeping the API up, that would be super :)
Got a good laugh from this one.<p>> GPT-NSFW is an N-gram model that was created using the same WebText dataset as GPT-2, but that is designed to generate NSFW text. The NSFW version of GPT-2 has shown great promise in generating NSFW text.<p><a href="https://thisaidoesnotexist.com/model/GPT-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" rel="nofollow">https://thisaidoesnotexist.com/model/GPT-NSFW/JTdCJTIyZGVmbi...</a>
I got:<p>> SpotifAI is a system that uses deep learning to automatically create playlists from user-submitted playlists. Its algorithm has been trained on millions of playlists from Spotify.<p>Which is pretty cool sounding and has a cool name.
Jesus is a fast and scalable language model trained on the Jesus dataset, which consists of over 4.7 billion words from the Bible. Jesus demonstrates state-of-the-art performance on several language modeling and conversational tasks.
Seems legit.<p>GPT-WESTWORLD is a large-scale, multilingual language model that generates fluent, realistic sentences from text in any language. It achieves this by incorporating a novel approach to language modeling and incorporating a new type of recurrent network, the Westworld.<p><a href="https://thisaidoesnotexist.com/model/GPT-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" rel="nofollow">https://thisaidoesnotexist.com/model/GPT-WESTWORLD/JTdCJTIyZ...</a>
Are these procedural or is there a list of pre-generated "AI"s next goes thru?<p>I got this as my third which seemed either prophetic or deterministic.<p>HackerNewsReplyGuy:<p>>from hackernews_response_guy import HackerNewsReplyGuy<p>>model = HackerNewsReplyGuy(1)<p>>model.predict_comments(comments, [u'comment_id'])
As someone who has trained around 60 GPT-2s, this is damn impressive work. It’s very hard to get consistent code quality when the training corpus is so small (as this one undoubtedly was).<p><a href="https://thisaidoesnotexist.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" rel="nofollow">https://thisaidoesnotexist.com/model/MozartNet/JTdCJTIyZGVmb...</a><p>The url scheme is interesting. I wonder what it base64 decodes to. If I were at a computer I’d check. It might be a complete representation of the inputs to the model, which is then cached. Which implies you might be able to fiddle with it to get specific outputs.
> AutoProfit is a reinforcement learning model that trains itself on a simulated trading environment. It is able to trade on its own and generate its own trading signals, outperforming a portfolio of human traders and making the most out of available information. AutoProfit is a model for trading stock, cryptocurrencies, and commodities in real time, generating trading strategies for itself. It uses an iterative training process, and has been tested on over 50 trading strategies.<p>Cool.
Very cool!<p><a href="https://github.com/thesephist/modelexicon" rel="nofollow">https://github.com/thesephist/modelexicon</a><p>Looks like it’s powered by GPT-J. My understanding is that GPT-J has comparable performance to OpenAI’s Curie model on many tasks (their second-best variant of GPT-3) but it’s an openly available model that you can run yourself if you have the resources.
Clicked into it, didn't read the description, and got an AI-based project that could perfectly hedge my fixed income portfolio. I won't lie, got a bit excited and then I realized what site I'd clicked on.<p>Very nifty! Is this your site?
This one gave me a good chuckle.<p>>TinderSwindler is a system developed by Facebook to analyze mobile phone location data in order to catch potential cheaters. TinderSwindler leverages Al technology to automatically identify relationships between people based on their movements over a period of time. TinderSwindler was released by Facebook in January 2018.
Some funny responses i've got:<p>Portal 3 spoilers:<p>GLaDOS is a character voiced by Ellen McLain that serves as the main antagonist of the Portal franchise. GLaDOS was originally a self-aware A.I. in the form of a computer that was built as a personality core for the Aperture Science Laboratories' mainframe. She is the main antagonist in the first game, Portal, and serves as a narrator for the second game, Portal 2. She is also the main antagonist of the third game, Portal 3, where she becomes the leader of the Aperture Science Resistance.<p>A semi-successful attempt at recursion:<p>thisaidoesnotexist is a tool that is able to generate fake images with high resemblance to real ones. This is achieved by using the GAN to generate the image, and then replacing the generated image with the real one.
I got <i>Timelord</i>:<p><i>Timelord is a self-supervised temporal model that learns a shared embedding of timestamped data. It is used as a pre-processing step in self-supervised training for a number of tasks such as semantic video segmentation and video captioning.</i><p>Now I want a library with that name<p>[1] Link to its description:
<a href="https://thisaidoesnotexist.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" rel="nofollow">https://thisaidoesnotexist.com/model/Timelord/JTdCJTIyZGVmbi...</a>
Skynet is an end-to-end speech recognition model. It is based on the Inception-v3 architecture and the Speech Transformer (Sphin) speech model. Its speech model was trained on a dataset of 30,000 hours of human speech, as well as speech recordings from the Switchboard corpus and the Fisher corpus. The model achieves 99.34% WER on the Switchboard-1.1 test set.
All this talk about Gateway Timeouts made me curious:<p>> Gateway Timeout is a deep learning-based anomaly detection system. It detects anomalies by learning the probability distribution of normal traffic and comparing it to traffic that does not match the normal distribution.
Congrats on this !<p>Hey OP, I work at HF, feel free to open an issue here <a href="https://github.com/huggingface/api-inference-community/issues" rel="nofollow">https://github.com/huggingface/api-inference-community/issue...</a> or contact api-enterprise@huggingface.co.<p>We've increased resources for you, and we'll check that things run as smoothly as they can.
This AI Does Not Exist (IDEA) is an AI system that can answer questions about itself. The system was created by the research team at the University of Cambridge and is based on the concept of an "AI mirror", which can be trained to look at itself and answer questions about its own existence.
Looks like we're currently getting a pre-defined set of 38 models: <a href="https://raw.githubusercontent.com/thesephist/modelexicon/main/models.json" rel="nofollow">https://raw.githubusercontent.com/thesephist/modelexicon/mai...</a>
HackerNewsReplyGuy is a bot for the Hacker News comment section. It consists of an encoder-decoder transformer model that is trained on the whole comment section. It has shown to be useful for spam detection and to reduce comment section noise.
this gave me weird dream last night;) pretty surreal<p>I iterally dreamed of some none nonsensical problem and in the dream I was like wait a second I have seen the nonsensical solution before (which happened to be one of the AI that doesn't exist)
It asked me for a model, so I naturally thought of female models and cars, decided upon "911" and get:
"911 is a dataset for 9/11 related tasks, including predicting the location of the first plane crash, the location of the second plane crash, and the location of the towers."<p>Thats not what I had in mind so it still needs a bit of work I think or at least the questions do. ;-)
You know now that I made friends with a homeless beggar I have no trouble making friends with a bot. Why not? Has some humanity breathed into them, like a book for instance, a book can be your friend. A kind old family friend who let me stay with her told me a long time ago just that when talking about a chest full of books, <i>these books are my friends</i>.