mstdn.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A general-purpose Mastodon server with a 500 character limit. All languages are welcome.

Administered by:

Server stats:

12K
active users

#transformermodel

0 posts0 participants0 posts today
MastodonBrice<p><a href="https://mstdn.social/tags/TransformerModel" class="mention hashtag" rel="tag">#<span>TransformerModel</span></a><br /><a href="https://mstdn.social/tags/NewYorkCity" class="mention hashtag" rel="tag">#<span>NewYorkCity</span></a><br /><a href="https://mstdn.social/tags/Brooklyn" class="mention hashtag" rel="tag">#<span>Brooklyn</span></a><br /><a href="https://mstdn.social/tags/ConeyIsland" class="mention hashtag" rel="tag">#<span>ConeyIsland</span></a><br /><a href="https://mstdn.social/tags/LunaPark" class="mention hashtag" rel="tag">#<span>LunaPark</span></a><br />@LunaPark<br /><a href="https://mstdn.social/tags/ElectroSpin" class="mention hashtag" rel="tag">#<span>ElectroSpin</span></a><br /><a href="https://mstdn.social/tags/WonderWheel" class="mention hashtag" rel="tag">#<span>WonderWheel</span></a><br /><a href="https://mstdn.social/tags/BriceDailyPhoto" class="mention hashtag" rel="tag">#<span>BriceDailyPhoto</span></a></p>
Dave Spector<p>Had perhaps the <a href="https://mastodon.social/tags/geekiest" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>geekiest</span></a> <a href="https://mastodon.social/tags/tshirt" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>tshirt</span></a> ever printed up ..because well, we are going to be living with this, for better or worse, for a while. <a href="https://mastodon.social/tags/TransformerModel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TransformerModel</span></a> <a href="https://mastodon.social/tags/AttentionIsAllYouNeed" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AttentionIsAllYouNeed</span></a> <a href="https://mastodon.social/tags/aihype" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aihype</span></a> <a href="https://mastodon.social/tags/AIpocalypse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIpocalypse</span></a></p>
Gareth Emslie 🇿🇦 🇪🇦 🇨🇭<p>The post discusses the impact of GPT-4, the most advanced version of a transformer model, on the development of Generative AI tools. These tools create content mimicking a particular style using a self-attention mechanism. The post also highlights the potential fears associated with such tools. <a href="https://blog.cloudflare.com/secure-generative-ai-applications/" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="ellipsis">blog.cloudflare.com/secure-gen</span><span class="invisible">erative-ai-applications/</span></a> <a href="https://hachyderm.io/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://hachyderm.io/tags/GPT4" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GPT4</span></a> <a href="https://hachyderm.io/tags/TransformerModel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TransformerModel</span></a> <a href="https://hachyderm.io/tags/softcorpremium" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>softcorpremium</span></a></p>
Arto Thurlin<p><a href="https://sigmoid.social/tags/ChatGPT" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ChatGPT</span></a> is all the rage. What's this all about and what's the wider context? This paper gives a nice and thorough survey of <a href="https://sigmoid.social/tags/TransformerModel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TransformerModel</span></a></p><p><a href="https://arxiv.org/abs/2302.07730" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2302.07730</span><span class="invisible"></span></a></p>
Elio Campitelli<p>Interesting preprint about image retrieval in image generation models. </p><p>They find that <a href="https://mastodon.social/tags/StableDiffusion" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>StableDiffusion</span></a> generates the same sofa 20% of the time when prompted with "Canvas Wall Art Print". </p><p>The problem seems to be that the training dataset has many repeated images from printshops.</p><p><a href="https://arxiv.org/pdf/2212.03860.pdf" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/pdf/2212.03860.pdf</span><span class="invisible"></span></a></p><p><a href="https://laion-aesthetic.datasette.io/laion-aesthetic-6pls/images?_search=Original+Oil+Painting+Canvas+Wall+Art+Print&amp;_sort=domain_id" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="ellipsis">laion-aesthetic.datasette.io/l</span><span class="invisible">aion-aesthetic-6pls/images?_search=Original+Oil+Painting+Canvas+Wall+Art+Print&amp;_sort=domain_id</span></a></p><p><a href="https://mastodon.social/tags/transformermodel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transformermodel</span></a> <a href="https://mastodon.social/tags/generativeart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>generativeart</span></a> <a href="https://mastodon.social/tags/aiart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aiart</span></a></p>
Jelle Zuidema<p>Interesting investigation of how close generated images (Stable D) are to training set images. </p><p>(I'd avoid terms like "stealing" and "blatantly copy" -- results speak for themselves).</p><p>RT: <span class="h-card"><a href="https://det.social/@HxxxKxxx" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>HxxxKxxx</span></a></span> <br>Do diffusion models create unique works of art, or are they stealing content directly from their training sets?</p><p>📑Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models</p><p>Via <span class="h-card"><a href="https://sigmoid.social/@akhaliq" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>akhaliq</span></a></span> <span class="h-card"><a href="https://sigmoid.social/@zentralwerkstatt" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>zentralwerkstatt</span></a></span> <br><a href="https://sigmoid.social/tags/transformermodel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transformermodel</span></a> <a href="https://sigmoid.social/tags/generativeart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>generativeart</span></a> <a href="https://sigmoid.social/tags/aiart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aiart</span></a> </p><p><a href="https://arxiv.org/abs/2212.03860" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2212.03860</span><span class="invisible"></span></a></p>
Harald Klinke<p>Do diffusion models create unique works of art, or are they stealing content directly from their training sets?</p><p>📑Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models</p><p>Via <span class="h-card"><a href="https://sigmoid.social/@akhaliq" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>akhaliq</span></a></span> <span class="h-card"><a href="https://sigmoid.social/@zentralwerkstatt" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>zentralwerkstatt</span></a></span> <br><a href="https://det.social/tags/transformermodel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transformermodel</span></a> <a href="https://det.social/tags/generativeart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>generativeart</span></a> <a href="https://det.social/tags/aiart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aiart</span></a> </p><p>&gt; <a href="https://arxiv.org/abs/2212.03860" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2212.03860</span><span class="invisible"></span></a></p>
Harald Klinke<p>If text-to-image models such as <a href="https://det.social/tags/dalle2" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>dalle2</span></a> can be thought of as searches on large amounts of image data, is it then theoretically possible, given the right input, to find/generate *exactly* one of the input images? </p><p><a href="https://det.social/tags/transformermodel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transformermodel</span></a> <a href="https://det.social/tags/GAN" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GAN</span></a> <a href="https://det.social/tags/generativeart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>generativeart</span></a> <a href="https://det.social/tags/AIart" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIart</span></a> <a href="https://det.social/tags/latentspace" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>latentspace</span></a> <span class="h-card"><a href="https://mastodon.social/@Bildoperationen" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>Bildoperationen</span></a></span> <span class="h-card"><a href="https://mastodon.social/@Quasimondo" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>Quasimondo</span></a></span></p>
IT News<p>AI model from OpenAI automatically recognizes speech and translates it to English - Enlarge (credit: Benj Edwards / Ars Technica) </p><p>On Wednesday, Op... - <a href="https://arstechnica.com/?p=1883524" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="">arstechnica.com/?p=1883524</span><span class="invisible"></span></a> <a href="https://schleuss.online/tags/speechrecognition" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>speechrecognition</span></a> <a href="https://schleuss.online/tags/transformermodel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transformermodel</span></a> <a href="https://schleuss.online/tags/machinelearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>machinelearning</span></a> <a href="https://schleuss.online/tags/speech" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>speech</span></a>-to-text <a href="https://schleuss.online/tags/whisper" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>whisper</span></a> <a href="https://schleuss.online/tags/biz" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>biz</span></a>&amp;it <a href="https://schleuss.online/tags/openai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>openai</span></a> <a href="https://schleuss.online/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ai</span></a></p>
IT News<p>New AI assistant can browse, search, and use web apps like a human - Enlarge / Still from a demo video showing ACT-1 performing a search on ... - <a href="https://arstechnica.com/?p=1881588" rel="nofollow noopener" target="_blank"><span class="invisible">https://</span><span class="">arstechnica.com/?p=1881588</span><span class="invisible"></span></a> <a href="https://schleuss.online/tags/transformermodel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>transformermodel</span></a> <a href="https://schleuss.online/tags/machinelearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>machinelearning</span></a> <a href="https://schleuss.online/tags/salesforce" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>salesforce</span></a> <a href="https://schleuss.online/tags/biz" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>biz</span></a>&amp;it <a href="https://schleuss.online/tags/act" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>act</span></a>-1 <a href="https://schleuss.online/tags/adept" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>adept</span></a> <a href="https://schleuss.online/tags/excel" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>excel</span></a> <a href="https://schleuss.online/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ai</span></a></p>